I*"-,
Birds, Traditional Coffee Plantations and Spatial complexity: TheDiversit y Puzzle
EuridiceLeyequie n Abarca
Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van Wageningen Universiteit, Prof. Dr. MJ. Kropff in het openbaar te verdedigen opvrijda g 21 april 2006 desnamiddag s te 16:00uu r in de Aula Leyequien Abarca, E. (2006) Birds, Traditional Coffee Plantations and Spatial Complexity: The Diversity Puzzle
Ph.D.-thesis, Department of Environmental Sciences, Sub-department ofNatur e Conservation, Resource Ecology Group, Wageningen University, theNetherlands , with summaries in English, Spanish and Dutch.
Keywords:Avia n diversity, Double-observer approach, Habitat loss, Habitat fragmentation, Low montane rain forest, Lower montane rain forest, Metacommunity, Mexico, Remote sensing, Species competition.
ISBN 90-850&-4\6-l Contents
Page 9 Chapter 1 General introduction Euridice Leyequien Abarca
Page 37 Chapter 2 Monitoring Neotropical birds: efficiency of amodifie d double- observer point counts approach versus mist-netting Euridice Leyequien, Willem F. deBoer, GarryBakker, Reinoud Vermoolen, Samuel Lopez de Aquino Submitted
Page 57 Chapter 3 Fragmentation and habitat loss: species richness responses in an avian metacommunity in Cuetzalan region, Mexico. Euridice Leyequien, Willem F. deBoer, VictorM. Toledo Submitted
Page 81 Chapter 4 Influence of body size on coexistence of bird species. Euridice Leyequien, Willem F. deBoer, Antoine Cleef Submitted Page99 Chapter 5 Linking species-environment relationships and multiple spatial scales in community ecology EuridiceLeyequien, W. Fred deBoer, Andrew K. Skidmore Submitted
Page 123 Chapter 6 Capturing the fugitive: applying remote sensing to terrestrial animal distribution and diversity. A review. Euridice Leyequien, Jochem Verrelst, Martijn Slot, Gabriela Schaepman- Strub,Ignas M.A. Heitkonig,Andrew Skidmore Accepted with minor revisions inInternationalJournal ofApplied Earth Observation and Geoinformation
Page 173 Chapter 7 Synthesis: Spatial complexity in the diversity puzzle: implications for conservation planning Euridice Leyequien Abarca
Page 207 Summary Resumen Samenvatting
Page227 Acknowledgments Curriculum vitae List of publications Affiliation of co-authors Page227 TSP Training and Supervision Plan of the Graduate School Production Ecology and Resource Conservation (PE&RC) if. \ ^* •"* Chapter 1
General introduction
Euridice Leyequien Abarca General introduction
Conserving the biological diversity on thisplane t is increasingly becoming a challenge due to the accelerated and increasingly loss of biodiversity amongst the contemporary conservationists. Ifw e want to avoid amas s extinction we cannot rely only on remnant patches ofnatura l areas to preserve biodiversity (Rosenzweig 2005,Toled o 2005). The previous protectionist paradigm where the isolation and protection of "natural areas"wa s placed above any social, economic, cultural or political consideration (Janzen 1986, Terborgh 199, Sanderson 2000) has led to the paradox of increasing loss of biodiversity.
Hence, conservation should also turn its efforts into the management and diversification of the anthropogenic matrix in which natural areas are embedded (Rosenzweig 2001, 2005). For example, in the case of tropical bird species, such linkages could be maintained by using the man-modified landscapes (e.g.;traditiona l shade coffee plantations) asthes e areas also harbour an ecological value ashabita t for wildlife species. To understand how humans can share the landscapes with other species we need tounderstan d which abiotic and biotic factors control species richness, and how these factors operate at multiple spatial scales, ultimately driving species persistence and coexistence.
Much arguments has been given about that independently ofth e scale in which the mechanism(s) that influence species richness patterns are measured, the results can be scaled up or down. However, several studies had proven that different processes are likely to determine species diversity at different spatial scales (Clarke and Lidgard 2000, Crawley and Harral 2001, Lennon et al.2001 ,Rahbe k and Graves 2001). It isbecomin g increasingly apparent that the factors best accounting for patterns of species diversity seems to be delimited by scale (Loreau 2000, Mouquet et al.2003 , Kneitel and Chase 2004, Borcard et al. 2004) where variables that best account for species richness on a local spatial scale may not be the same as those accounting for richness at regional spatial scales (Willis and Whittaker 2002).A hierarchical approach appears to be more appropriate for accurately modelling species richness where processes can be nested according to the scales (Willis
11 and Wittaker 2002). Thus,th e integration of scaling effects in studies of species- environment interactions and community composition is aprerequisit e for aiming at proper conservation strategies.
Besides, one of the major problems for species conservation isth e loss of habitat, and landscape fragmentation. In Latin America natural ecosystems are suffering from an accelerated deforestation, and there is ahig h rate of conversion to monocultures and grazing lands that has direct effects onNeotropica l avian communities (i.e.;resident and migrants). The ecological impacts of both habitat loss (i.e.;reductio n of total habitat area) and landscape fragmentation (i.e.;configuratio n changes in the landscape) onpattern s and dynamics are the disruption of important ecological processes for species persistence (e.g.; mortality, re-colonisation, orreproductiv e rates) (Bender etal. 1998,Fahri g 1997, 2001, Fahrig and Merriam 1994, Stephens etal. 2003 , Villard etal. 1995,Wiegan d etal. 2005) . Specifically, habitat loss affects (among others) the breeding and dispersal success (Belisle etal. 2001, Kurki etal. 2000 ,Wit h and King 1999),trophi c chain length (Komonen etal. 2000), species interactions (Taylor and Merrian 1995), species-specific extinction thresholds (With and Crist 1995), and has often been described as only having negative effects on species diversity.
In addition, habitat fragmentation, which leads to an increasing number and isolation of patches while decreasing patch sizes, changes the spatial structure of landscapes. The effects of fragmentation had been accounted aseithe r positive or negative in a number of studies (e.g.; McGarigal and McComb 1995,Bende r et al. 1998,Kremsate r and Bunnell 1999, Trzcinski etal. 1999). Besides, both positive and negative effects of fragmentations may be species-dependent, or landscape configuration characteristics such aspatch edge which could have positive effects on abundance or distribution of species (Kremsater and Bunnell 1999).Despit e the important effects that habitat loss and fragmentation may cause to species survival and maintenance, aconsiderabl e amount of research had lead to small room for generalisations and often show ambiguous results (for a review see Fahrig 2003). Most researchers lump both the loss ofhabita t and the breaking apart of habitat into the
12 single concept of habitat fragmentation, producing unclear results about the magnitude and direction of its individual effects on species diversity (Fahrig 2003).W e argue that generalisation ispossibl e only for the separate components of loss and breaking apart of habitat, and that the relative influence of eachproces s should be accounted separately if we are to design proper conservation strategies.
Moreover, local communities are often connected by dispersal at the landscape or regional scale (Wilson 1992, Holt 1993,Hubbel l 2001, Mouquet and Loreau 2002). Local extinctions and colonisations can be influenced by interspecific interactions such as competition. Understanding what determines interspecific competition and its strength will give us insight into the forces shaping the community patterns. Body mass is an easily determined characteristic of a species that probably influences competition interactions (Bowers and Brown 1982, Brown and Maurer 1986,Frenc h and Smith 2005,Gotell i and Ellison 2002) and could provide areasonabl e simple variable to assess competition strength.
Another important issue in conservation biology isth e monitoring of community trends to provide reliable information of species diversity and their status, for fast and efficient identification of conservation priorities. Species richness lists ofbir d communities are crucial to ecological research, conservation plans and environmental impact assessments. The problem arises when using different techniques to acquire the species information as this may lead to dissimilitude in the assessment of species richness, and therefore producing unreliable data about bird communities. Moreover, the costs i.e.; time and money related to the implementation of different techniques is inmos t of the cases a strong criterion in the decision ofwhic h technique to use. This decision may have important consequences inth e reliability and completeness of the species richness assessments, thus it isrelevan t to monitoring programs to consider the biases,cost s and efficiency of thepotential methods to be used. Besides the field monitoring of species, a constantly developing technology, i.e.; remote sensing had brought the possibility to acquire data about the Earth's surface with high spectral, spatial and temporal resolution, providing a fast and efficient way of
13 obtaining species data over time, with spatially continuous coverage. The potential of this remotely sensed information opens vast possibilities for conservation planning and monitoring of faunal species (Leyequien etal. accepted) .
Why birds and traditional shade coffee plantations?
Neotropical avifauna comprises a high diversity and is worldwide recognised as threatened by habitat loss and fragmentation due to human activities (Winker 1998). In comparison to Neartic birds,ther e is still a gap in the knowledge about Neotropical birds, and this a serious concern in conservation biology. Several studies found that anumbe r of resident and migratory birds inth e Neotropics are dependent on agroecosystems (e.g.; traditional shade coffee plantations) for food and shelter resources aswel l as for connecting corridors (Calvo and Blake 1998, Wunderle 1999, Moguel and Toledo 1999, Van der Voort and Greenberg 1999). Thus,traditiona l shade coffee plantations may function as important areas for anumbe r ofNeotropica l resident and migratory bird species (Pimentel etal. 1992 , Perfecto etal. 1996 , Mehta and Leuschner 1997, Moguel and Toledo 1999).
However, the current trend innorther n Latin America (Colombia, Panama, Costa Rica, El Salvador, Guatemala, Nicaragua, Honduras and Mexico) ist o reduce shade cover plantations and convert them into reduced or non-shaded coffee plantations. It has been reported for northern Latin America, that about 41% of the 2.7 million ha devoted to coffee production have been converted to reduced ornon-shad e coffee plantations (Rice and Ward 1996). This trend could have astounding implications for the loss of biodiversity including bird species; in addition to the ecological services that these agroecosystems provide at local, regional and global level such as maintaining the hydrological balance, improving the
soil quality, increasing the forest cover, or improving the C02 balance (Moguel and Toledo 1999). However, despite the importance of agroecosystems for biodiversity conservation (Liung etal. 2001) ,ther e has been more attention on biological diversity in undisturbed ecosystems.
14 Aimo fthi s study
What controls species richness? The issue of scale complicates the quest for finding the answers of thispivota l question for ecologists and other scientists. Processes and patterns that influence species persistence and coexistence are dependent on the spatial scale that we choose to study (Rosenzweig 1999). Moreover, how these spatial patterns are approached in ecological studies could have significant consequences for conservation planning. In addition to the issue of scale, finding optimal solutions for the design of large-scale monitoring is essential for biodiversity conservation.
The aim of this thesis ist o give insight into the forces shaping ecological communities, addressing the issue of scale, topu t forward effective monitoring techniques for Neotropical avifauna, and to give insight into the conservation implications of traditional shade coffee plantations for the conservation of biological diversity, inthi s case the Neotropical avifauna.
Site description
Mexico supports about 1,060 bird species, i.e.; 11% ofth e total ofworld' s bird species, and ranks fifth in the world for the number of endemic bird areas,wit h 10% ofth e bird species endemic to the country. Because of its strategic position atth eNearcti c and Neotropical biogeographic boundary, Mexico plays a crucial role in the conservation of migratory birds. The study area is located in theNort h Eastern mountain range of Puebla,withi n the Cuetzalan region, Mexico, and covers an area of- 151,368h a Puebla is situated on the east migratory paths ofNeartic-Neotropica l birds that winter or refuel inthi s region (Figure 1).
15 -40 -20
-60-40-20
Figure 1.Neotropical-Neartic birdmigration routes, passing from Canadaand North America to the neotropics (Mexico toSouth America).
Moreover, the Cuetzalan region ispredominantl y a coffee agro-ecological zone, which covers an altitudinal range between 300t o 1200m . The major agricultural system is traditional polyculture (traditional shade coffee plantation),whil e other land covers, such as seasonal crops (i.e.;maiz e and beans) and grazing lands are also present inth e area. Forested areas aretypicall y remnants oftropica l evergreen forest with low montane rain forest in the higher areas (between 700 and 2500 m) and lowland rain forest in the lower areas (200-700 m) (Figure 2). The mean annual rainfall is 3,500m m without a defined dry season; and the mean annual temperature is22-26°C .
16 CeO £ 3 (/) g>
N 0 O
645000 650000
JO km
6*5000 650000 6B5000 690000
Figure 2. Location of the research area and land coversAGL = Agricultural lands, CP= Coffeeplantations, F =Lower tropical evergreenforest, SV =Secondary vegetation, UA = Urbanareas, WB = Waterbodies, CI= Clouds; the thick black line is thearea of influence of the Cooperative Tosepan Titataniskewhere all the research samples occur.We performed a supervised classification using themaximum likelihood classifiermethod, includingprior probabilities togenerate the land cover map.
The studies described inthi s thesis were conducted within traditional shade coffee agroecosystem,an d enclosed eight localities (i.e.;Acaxiloco ,Yohualichan , Pinahuistan, Reyeshogpan, Limonco,Atalpan , Monte Alto and Tozan) situated inth e municipalities of Cuetzalan del Progreso and Jonotla.
17 Coffee plantations inMexic o
There are five main coffee agricultural systems in Mexico, differentiated according to management level, vegetation composition and structural complexity (Fuentes-Flores 1979, Nolasco 1985) (Figure 3).
i_ ^
Figure 3. Thefive coffee agroecosystems ofMexico, illustrating thestructural andfloristic complexity of vegetation and height of canopy (source: Gobbi2000).
18 Traditionalrustic or "mountain". I n this agro-ecosystem, the herbaceous stratum is removed from the original vegetation cover and replaced by coffee plants, while the shrub and arboreal strata remain. This type of agro-ecosystem has the lowest impact on the original vegetation cover and requires low labour-material inputs;howeve r also produces generally low yields. In Mexico, this type of agro-ecosystem is located in relatively isolated areas managed by indigenous communities, which introduced coffee plants in their native forest (Moguel and Toledo 1999).
Traditionalshade coffeeplantations (polyculture). This agro-ecosystem, as in the traditional rustic, coffee plants are introduced below the canopy. However, certain original canopy species are removed, others maintained, or domesticated species introduced (e.g.; Inga sp.,Pimenta dioica, Cedrelaodorata, Swietenia macrophylld) as sources for shade, food, medicine or raw materials for market or local subsistence. While these species provide arboreal cover, coffee and other cash crops (citrus fruits, bananas, and other fruit trees), they grow inth e shadow provided by the arboreal stratum.
This type of management increases the structural complexity of the vegetation and the species richness, asbot h wild and domesticated species co-exist. It has been estimated that this agro-ecosystem covers almost 50%o fth e Mexican coffee lands. Only low labour and material inputs (although some use agrochemicals), yet yields can be comparable to those of commercial "modern" systems (Moguel and Toledo 1999).
Commercialpolycultures. There is a complete removal of the original canopy trees in this agricultural system. Introduced tree species shade the coffee shrubs. These shade trees provide some economic utility and/or add nitrogen to the soil. Some of themos t common species are rubber trees (Castilla elastica), old spice {Pimentadioica), cedar{Cedrela odorata),jiniqui l {Ingaspp.) ,an d colorin {Erythrinaspp. ) (Moguel and Toledo 1999). Other cash crops are grown under the canopy such as citrus, banana or other fruit trees.
19 This system ismarket-oriented , has high labour and material inputs (agrochemicals are regularly used) and is characterised by ahighe r coffee yield. Species diversity is however considerably lower than inth e traditional rustic or traditional polyculture systems (Moguel and Toledo 1999).
Shaded monoculture. This is a"modern " cultivation where leguminous trees {Ingaspp ) almost exclusively provide shade to coffee shrubs. The result is an almost a mono-species canopy. The use of agrochemical products iscompulsory , and it is exclusively market- oriented (Moguel and Toledo 1999).
Unshadedmonoculture. This agricultural system has no tree stratum and the coffee shrubs are exposed to direct sunlight. Its management has lost the agroforestry character present in the other systems mentioned before. This specialised plantation requires high labour and material inputs (e.g.;chemica l fertilizers andpesticides) . The highest yields are reported under this system (Moguel and Toledo 1999).
Economic importance ofcoffe e plantations
Worldwide, one ofth e most traded commodities is coffee, producing important economic benefits for developing countries. The International Coffee Council (2001) estimated that over 125millio n ofpeopl e are dependent on coffee for their income. Only in northern Latin America 3.1 million hectares are coffee lands generating 34%o f the worldwide production (Perfecto and Ambrecht 2002).
Coffeeproduction inMexico. Mexico isranke d worldwide as number eight in coffee volume production, after Brazil, Vietnam, Colombia, Indonesia, Peru, India, and Papua Guinea (Tablel), and seventh inyiel d performance (ICO 2005). Coffee production, in Mexico, is situated in 12state s from which five are the major coffee producers (i.e.; Chiapas,Oaxaca , Veracruz, Puebla and Guerrero).Coffe e lands cover approximately 806,000 hectares (based on statistics ofNovembe r 2005, SAGARPA 2005) distributed among 250,000 small-scale producers from which 185,000belon g to indigenous groups
20 from the states of Chiapas, Oaxaca, Puebla or Guerrero (Perez-Grovas et al.2001) .I n Mexico, coffee represents one ofth e most important agricultural sources of foreign exchange (18%o ftota l export) generating annually 500,000 direct and 3,000,000 indirect jobs. In addition, Mexico is the world's largest producer of organic coffee, accounting for about 20% ofth e global market (Perez-Grovas etal. 2001) .
Table 1. Coffeeexported by the top ten coffee-producing countries, bothArabica and/or Robusta inpercentage of thousand of bags of 60kg. ^Thousandsof bagsof60 kg.A = Arabica andR = Robusta. Modifiedfrom ICO (2005).
Country Coffee type Oct-2005 Nov2004-Oct2005 Oct-2004 Nov2003-Oct2004 Brazil A/R 34.8 30.4 36.9 28.6 Vietnam R 15.3 15.6 15.9 16.7 Colombia A 12.2 12.3 12.8 11.3 Indonesia R/A 7.4 6.8 6.5 5.9 Peru A 7.3 3.2 6.9 3.1 India A/R 3.2 3.1 2.6 4.2 Papua New Guinea A/R 2.5 1.3 1.3 1.2 Mexico A 2.1 2.1 1.4 2.7 Worldwide Total 6164390 88451808 6880365 89143636
Social background
The Regional Agricultural Cooperation "Tosepan Titataniske" is an indigenous organisation that groups 5,800coffe e farmers (i.e.;Nahuat s and Totonacos) from the north eastern mountain range of Puebla,Mexico . The cooperation is engaged with the social and economic welfare ofth e coffee farmers and their families, and has ahig h interest in the conservation of their natural resources. Their membership is oriented to small producers, labourers,house landladies and craftsmen. The cooperative produce organic coffee, and there isa growing number of coffee plantations converting to organic production. Moreover, the implementation of a floristic diversification program (native trees and plants) on their coffee plantations is currently carried outb y the coffee farmers to avoid erosion and nutrient enrichment, to provide suitable habitat for wildlife species andt o supply with alternative agricultural products for self consumption and local and global markets
21 (Tosepan Titataniske 2005). The Tosepan Titataniske area of influence encloses a large area of out the study region. Therefore this research wasjointl y developed with the Tosepan Titataniske toprovid e ecological information ofth e avifauna present inthei r coffee lands, as well as guidelines for biodiversity conservation and ecotourism programmes.
The importance of management actions carried out by indigenous groups, especially small farmers, has been pointed out by the United Nations University (UNU). The UNU recommended studies to examine the relationships between population and environmental change, and stimulated the participation of small farmers in the conservation and sustainable use ofnatura l resources, especially, biodiversity. Hence,ther e is a need to analyze, model, and assess the ecological implications of these coffee agroecosystems.
Ecological importance of coffee plantations: bird responses
Mexico comprises more than 12% ofth e worldwide biota and isplace d among the seven biologically most diverse countries in the world (Ramamoorthy etal. 1993). The10 %o f its biodiversity is located in areas managed by peasants that in many cases belong to an indigenous group (Ramamoorthy et al. 1993, Challenger 1998). Coffee plantations represent one ofth e most important land covers among indigenous communities (Perez- Grovas etal. 2001) . Approximately 806,000 hectares are under coffee plantations, and it has been estimated that 60-70% of the coffee areas are managed under traditional shade plantations. Moreover, 9% ofthes e coffee areas overlap (or are near) regions with high biological diversity and endemicity, which are regarded as crucial to biodiversity conservation in Mexico (Moguel and Toledo 1999) (Figure4) .
22 PriorityArea sfo r Conservation Coffee Regions
Figure 4.Priority areasfor conservation inMexico (basedon CONABIO2005) and coffee regions.
Shade coffee systems have been recognised in Latin America as important and potential areas for biodiversity conservation (Pimentel etal. 1992,Perfect o etal. 1996,Meht a and Leuschner 1997,Mogue l and Toledo 1999, Gobbi 2000). Several studies have found that shade coffee plantations, especially the traditional rustic and the traditional polyculture, represent an important habitat for wildlife (Calvo and Blake 1998, Moguel 1996,Va n der Voort and Greenberg 1999, Wunderle 1999). Studies ofbiologica l diversity in different types of coffee plantations have been conducted for diverse taxa, e.g.; arthropods (Torres 1984, Lopez-Mendez 1985,Moro n 1985,Ibarra-Nune z 1990, Stork and Brendell 1990, Nestel et al.1993, Perfecto and Snelling 1995,Llorente-Bousquet s etal. 1996,Perfect o et al. 1997,Estrad a etal. 1998,Ibarra-Nune z and Garcia-Ballinas 1998,Ma s 1999, Johnson 2000, Molina 2000, Sadeghian 2000, Perfecto etal. 2003 , Pineda etal. 2005) , mammals (Estrada etal. 1993, 1994,Gallin a etal. 1996,Estrad a etal. 1998,Pined a etal. 2005) , amphibians and reptiles (Rendon-Rojas 1994, Lenart etal. 1997, Pineda etal. 2005) , macrofauna (Fragoso etal. 1993), and birds (Terborgh and Weske 1969,Aguilar-Orti z 1982, Borrero 1986,Robbin s etal. 1992,Wunderl e and Widel993, Vennini 1994, Thiollay 1995, Greenberg 1996,Perfect o etal. 1996,Wunderl e and Latta 1994, 1996,Greenber g et
23 al. 1997, Calvo and Blake 1998, Petit etal. 1999,Va n der Voort and Greenberg 1999, Dietsch 2000, Johnson 2000, Tejeda-Cruz and Sutherland 2004). The general findings of the mentioned studies have shown that traditional shade coffee plantations hold high species richness for broad taxa compared to non-shade coffee plantations and other agricultural systems. In general, though the species composition differs from that of natural forests, many forest-dependent species or species that depend on dense high canopy are still found in traditional shade coffee plantations. Moreover, the species richness in these agroecosystems, in some cases, is nearly equal or even higher to the species richness of natural forests.
Avifauna
Avian studies in Latin America demonstrated ahig h species diversity and elevated bird densities in traditional shade coffee plantations, and found similar levels ofbir d species richness compared to natural forest (Aguilar-Ortiz 1982,Robbin s etal. 1992,Wunderl e and Wide 1993,Vennin i 1994,Wunderl e and Latta 1994, 1996,Greenber g etal. 1997,Va n der Voort and Greenberg 1999, Dietsch 2000,Johnso n 2000). Species composition in traditional shade coffee plantations may vary from natural forest, because forest-edge and second-growth species contribute significantly to the high diversity of birds in coffee plantations (Greenberg etal. 1997). It has been found in several studies that a large percentage of thebird s in shade coffee plantations are omnivorous andpartia l nectarivorous (Wunderle and Latta 1996,Greenber g etal. 1997,Tejeda-Cru z and Sutherland 2004). Traditional shade coffee plantations have been recognised as an important habitat for migratory birds, which can be found even in higher densities than innatura l forests (Borrero 1986,Greenber g etal. 1997). Moreover, most affected by habitat modifications are large frugivorous and insectivorous species ofth e canopy and low understory, or terrestrial interior forest specialists (Thiollay 1995).I n addition, it has been suggested that traditional shade coffee plantations are an important refuge area for birds during the dry season when energetic demands are high and when food is scarce in other habitat types (Wunderle and Latta, 1994,Johnson , 2000). Wunderle and Latta (1998), Greenberg etal.
24 (1997) and Johnson (2000) stated that the tree canopy species in traditional shade coffee plantations could provide important food sources for nectarivorous and insectivorous birds.
The role ofth e avifauna in the ecosystem functioning of this agroecosystem (i.e.; traditional shade coffee plantation) had received significant attention (for arevie w see Perfecto and Armbrecht 2003). Itha s been suggested that the structural complexity and floristic diversity in traditional shade coffee plantations support ahig h density and diversity ofpredators , which are responsible for areduce d number ofpest s (arthropods) in plantations (Ibarra- Nunez 1990,Perfect o and Castineiras 1998). Many bird species registered in shade coffee plantations are either insectivorous or omnivorous, with arthropods as the majority of their diet. Additionally, experimental enclosure studies have demonstrated that birds often remove a great portion ofth e arthropod populations, inparticula r large herbivorous arthropods (Holmes etal. 1979, Gradwohl and Greenberg 1982, Moore and Yong 1991, Bock etal. 1992, Marquis and Whelan 1994, Perfecto etal. 2004) , suggesting that the effect of birds is quite equally distributed across the ecological and taxonomic arthropod groups (Greenberg etal. 2000) . Based on the latter information, Perfecto and Armbrecht (2003) suggested that diversely shaded coffee plantations could be more resistant to pest outbreaks than unshaded monocultures because of the higher diversity and density of insectivorous birds in these plantations.
Outline of this thesis
Chapter 2 examines methodological aspects of avifauna diversity assessments and evaluates the relative biases, costs and efficiency of field monitoring techniques for avifauna. Chapter 3 examines the influence ofhabita t loss and landscape fragmentation on the species richness ofresident birds. The analysis focuses on the unimodal and linear response of species richness to habitat loss and fragmentation, respectively. Chapter 4 investigates the relative influence ofbod y mass on the strength of interspecific competition between birds from similar food guild. Chapter 5o fthi s study deals with how bird species-environmental relationships operate at multiple spatial scales. The relative importance of each scale and
25 the confounding effects between the scales on bird species richness and bird abundance was analysed. Chapter 6 gives an overview of studies in which remote sensing was used to assess faunal biological diversity, considering the current state of the artwit h regard to the development ofproxie s and methodologies for estimating the diversity in animals, and explores its challenges for contributing to conservation and sustainable use of biodiversity. In Chapter 7 finally, Isynthesis e the findings ofthes e studies, and illustrate the conservation andmanagement implications for Neotropical birds in shade coffee plantations.
26 References
Aguilar-Ortiz, F. 1982. Estudio ecologico de las aves del cafetal, in Estudios Ecologicos en el Agroecosistema Cafetalero, Jimenez-Avila, E and Gomez-Pompa, A. (eds.) INIREB, Xalapa, Veracruz, Mexico. Belisle, M., Desrochers, A. and Fortin, M.J. 2001.Influenc e of forest cover on the movements of forest birds:a homing experiment. Ecology 82: 1893-1904. Bender, D.J., Teschendorf, L. and Fahrig, L.1998. Evaluation ofpatch isolation metrics for predicting animal movement in binary landscapes. Landscape Ecology 15: 131- 143. Bock, C.E., Bock, J.H., and. Grant, M.C. 1992.Effect s ofbir d predation on grasshopper densities in Arizona grasslands. Ecology, 73:1706-1717. Borrero, H. 1986.L a substitution de cafetales de sombrio por caturrales y su efecto negativo sobre la fauna devertebrados . Caldasia 15: 725-7326. Borcard, D., Legendere, P.,Avois-Jacque t C, Tuomisto, H. 2004. Dissecting the spatial structure of ecological data atmultipl e scales. Ecology 85(7):1826-1832 . Bowers, M.A. and Brown, J.H. 1982. Body size and coexistence in desert rodents: chance or community structure? Ecology 63(2): 391-400. Brown, J.H. and Maurer, B.A. 1986.Bod y size, ecological dominance and Cope's rule. Nature 324:248-250 . Calvo, L. and Blake, J. 1998.Bir d diversity and abundance on two different shade coffee plantations in Guatemala. Birds Conservation International 8: 297-308. Challenger, A. 1998. Utilization y conservation de los ecosistemas terrestres en Mexico. Pasado,present e y future Comision Nacional para el Conocimiento y Uso de la Biodiversidad. Universidad Nacional Autonoma de Mexico, Agrupacion Sierra Madre, S.C., Mexico, D.F. Dietsch, T.V. 2000. Assessing the conservation value of shade-grown coffee: a biological perspective using Neotropical birds. Endangered Species Update 1750-72 [Electronic version].Retrieve d January 2, 2006, from http://www.umich.edu/- esupdatel
27 Estrada, A. 1994.Non-flyin g mammals and landscape changes in the tropical rain forest region of Los Tuxtlas. Ecography 17:229-241 . Estrada, A., Coates-Estrada, R. and Memt, D. 1993.Ba t species richness and abundance in tropical rain forest fragments and in agricultural habitats in Los Tuxtlas,Mexico . Ecography, 16:309-318. Estrada, A., Coates-Estrada, R., Dadda, A.A., Cammarano, P.1998 . Dung and carrion beetles in tropical rain forest fragments and agricultural habitats at Los Tuxlas, Mexico. Journal of Tropical Ecology 14:577-593 . Fahrig, L. and Merriam, G. 1994. Conservation of fragmented populations. Conservation Biology 8: 50-59. Fahrig, L. 1997.Relativ e effects of habitat loss and fragmentation on population extinction. Journal of Wildlife Management 61: 603-610. Fahrig, L. 2001.Ho w much habitat is enough? Biological Conservation. 100:65-74 . French, A.R. and Smith, T.B. 2005. Importance of body size in determining dominance hierarchies among diverse tropical frugivores. Biotropica 37(1): 96-101. Fragoso, C, Brown, G.G., Patron, J.C., Blanchart, E.,Lavelle , P., Pashanasi, B., Senapati, B. and Kumar, T. 1997.Agricultura l intensification, soil biodiversity and agroecosystem function in the tropics: the role of earthworms. Applied Soil Ecology 6: 17-35. Fuentes-Flores, R. 1979. Coffee production systems in Mexico. In: De las Salas, F. (ed.) Workshop on agroforestry systems in Latin America. Centro Agronomico Tropical de Investigation yEnsenanza , Turrialba, Costa Rica. Gradwohl, J. and Greenberg, R. 1982.Th e effect of a single species of avian predator on the arthropods of aerial leaf litter. Ecology, 63: 581-583. Gallina, S.E., Mandujano, S. and Gonzalez-Romero, A. 1996.Conservatio n of mammalian biodiversity in coffee plantations ofCentra l Veracruz, Mexico. Agroforestry Systems 33: 13-27. Gobbi, J.A . 2000. Is biodiversity-friendly coffee financially viable? An analysis of five different coffee production systems in western El Salvador. Ecological Economics 33:267-281.
28 Gotelli,N.J . and Ellison, A.M. 2002.Assembl y rules for New England ant assemblages. Oikos 99:591-599 . Greenberg, R. 1996. Bird populations in forest patches in Southern Mexico. In: Schelhas, J. and Greenberg, R. (eds.). Forest Patches in Tropical Landscapes. Island Press, Washington, DC. Greenberg, R., Bichier P. and Sterling J. 1997.Bir d populations in rustic and planted shade coffee plantations of Eastern Chiapas, Mexico. Biotropica 29: 501-514. Greenberg, R., Bichier, P. and CruzAngon, A. 2000. The impact of avian insectivory on arthropods and leaf damage in some Guatemalan coffee plantations. Ecology81 : 1750-1755. Holmes, R.T., Schultz, J.C. andNothnagle , P. 1979. Bird predation on forest insects: an exclosure experiment. Science 206:462463 . Holt, R. D. 1993.Ecolog y atth e mesoscale: the influence ofregiona l processes on local communities. In: Ricklefs, R. E. and Schluter, D. (eds.). Species diversity in ecological communities: historical and geographical perspectives. University of Chicago Press, Chicago. Hubbell, S. P. 2001.Th e unified neutral theory ofbiodiversit y and biogeography. Princeton University Press, Princeton, N.J. Ibarra- Nunez, G. 1990. Los artropodos asociados a cafetos enu n cafetal mixto del Soconusco, Chiapas,Mexico ,Varieda d y abundancia. Folia Entomologica Mexicana 79: 207-231. Ibarra-Nunez G. and Garcia-Ballinas J.A. 1998.Diversida d detre s familias de aranas tejedoras (Araneae: Araneidae, Tetragnathidae, Theridiiae) en cafetales del soconusco, Chiapas, Mexico. Folia Entomologica Mexicana 102: 11-20. International Coffee Organization. 2005.Export s by exporting countries to all destinations [Electronic version].Retrieve d January 2,2006 , from http://www.ico.org/prices/ml .htm Johnson, M.D. 2000.Effect s of shade-tree species and crop structure on the winter arthropod and bird communities in aJamaica n shade coffee plantation. Biotropica 32: 133-145.
29 Kneitel, J.M. and Chase, J.M. 2004. Trade-offs in community ecology: linking spatial scales and species coexistence. Ecological Letters 7: 69-80. Komonen, A., Penttilae, R., Lindgren, M. and Hanski, I. 2000. Forest fragmentation truncates a food chain base don an old-growth forest bracket fungus. Oikos 90: 119-126. Kremsater, L., and Bunnell, F. L. 1999.Edg e effects: theory, evidence and implications to management ofwester n North American forests. In: Rochelle, J. A., Lehmann, L. A. and Wisniewski, J. (eds.) Forest fragmentation: wildlife and management implications. Brill, Boston, Massachusetts, USA. Kurki, S.,Nikula , A., Helle,P . and Linden,H . 2000.Landscap e fragmentation and forest composition effects on grouse breeding success in boreal forests. Ecology 81: 1985-1997. Lenart, L.A., Powell, R., Parmerlee, J.S.Jr. , Lathrop, A. and Smith, D.D. 1997.Anolin e diversity in three differently altered habitats in the Sierra de Barouco, Republica Dominicana. Hispaniola Biotropica 29: 117-123. Leyequien, E.,Verrelst , J., Slot, M., Schaepman-Strub, G., Heitkonig, I.M.A. and Skidmore, A. (accepted inJAG). Capturing the fugitive: applying remote sensing to terrestrial animal distribution and diversity. A review. Liang, L., Stocking, M., Brookfield, H. and Jansky, L. 2001.Biodiversit y conservation through agrodiversity. Global Environmental Change. 11: 97-101. Llorente-Bousquets, J.A., Martinez, I., Vargas-Fernandez, I. and Soberon-Mainero, J. 1996. Papilionoidea (Lepidoptera). In: Lorente-Bousquets, J., Garcia-Aldrete, A. and Gonzalez-Soriano, E. (eds.).Biodiversidad , taxonomia y biogeografia de artropodos en Mexico, Mexico, D.F. Lopez-Mendez, J.A. 1989. La Actividad y la Depredation de Artropodos Asociados al Cultivo del Cafe-Cacao por dos Hormigas Ponerinas (Hymenoptera: Formicidae) en el municipio de Tuxtla Chico, Chiapas, B.Sc. thesis,Universida d Autonoma de Chiapas, Mexico. Loreau, M. 2000. Are communities saturated? On the relationship between a, Pan dy diversity. Ecological Letters 3:73-76 .
30 McGarigal, K. and McComb,W.C. 1995.Relationship s between landscape structure and breeding birds in the Oregon Coast Range.Ecologica l Monographs 65: 235-260. Marquis,R.J . and Whelan, J.C. 1994.Insectivorou s birds increase growth of white oak through consumption of leaf-chewing insects. Ecology, 75:2007-2014 . Mas, A.H. 1999. Butterfly diversity and the certification of shade coffee in Chiapas, Mexico, M.Sc. Thesis,Universit y of Michigan, Ann Arbor, Michigan. Mehta,N.G . and Leuschner, W.A. 1997.Financia l and economic analyses of agroforestry systems and a commercial timber plantation in the La Amistad Biosphere Reserve, Costa Rica. Agroforestry Systems 37: 175-185. Moguel, P. 1996.Biodiversida d y cultivos agroindustriales: el caso del cafe. Technical report. Instituto de Ecologia, Universidad Nacional Autonoma de Mexico, Mexico, D.F. Moguer, P. and Toledo, V.M. 1998.Biodiversit y conservation intraditiona l coffee systems of Mexico. Conservation Biologyl3(l): 11-21. Molina, J. 2000. Diversidad de escarabajos coprofagos (Scarabaeidae: Scarabaeinae) en matrices de la zona cafetera (Quindio-Colombia), in Memorias Foro Internacional Cafe y Biodiversidad, Federation Nacional de Cafeteros de Colombia (ed.), Chinchinil, Colombia. Moore,F.R . and Yong, W. 1991. Evidence of food-based competition among passerine migrants during stopover. Behavioral Ecology and Sociobiology 28:85-90 . Moron, M.A. and Lopez-Mendez, J.A. 1985.Analisi s de la entomofauna necrofila de un cafetal en el Soconusco, Chiapas, Mexico. Folia Entomologica Mexicana 63: 47- 59. Mouquet, N. and Loreau, M. 2002. Coexistence in metacommunities: the regional similarity hypothesis. American Naturalist 159:420-426 . Mouquet, N., Munguia, P.,Kneitel , J.M. and Miller T.E. 2003. Community assembly time and the relationship between local and regional species richness. Oikos 103:618 - 626.
31 Nestel, D., Dickschen, F. and Altieri, M.A. 1993.Diversit y patterns of soil macro- coleopterans in Mexican shaded and unshaded coffee agroecosystems: an indication of habitat perturbation. Biodiversity and Conservation 2: 7-78. Nolasco, M. 1985.Caf e y Sociedad en Mexico, Centro de ecodesarro, Mexico City. In: Oxfam, 2002. Bitter Coffee: How the Poor are Paying for the Slump in Coffee Prices. Perez-Grovas, V., Trejo, C. and Burstein, J. 2001.Cas e Study of the Coffee Sector in Mexico (Research Paper for Oxfam Coffee Report "Mugged"). Perfecto, I. and Castineiras, A. 1998.Deploymen t ofth e predaceous ants and their conservation in agroecosystems. In: Barbosa, P. (ed.) 1998. Conservation Biological Control. Academic Press,Washington , D.C. Perfecto I. and Snelling R. 1995. Biodiversity and the transformation of a tropical agroecosystem: ants in coffee plantations. Ecological Applications 5: 1084-1097. Perfecto, I., Rice, R.A., Greenberg, R. and Van der Voort, M.E. 1996. Shade coffee: a disappearing refuge for biodiversity. BioScience 46: 598-608. Perfecto, I., Vandermeer, J.H., Hanson, P. and Cartin, V. 1997.Arthropo d biodiversity loss and the transformation of a tropical agro-ecosystem. Biodiversity and Conservation 6: 935-945. Perfecto, I.an d Ambrecht, I. 2002. The coffee agroecosystems in the Neotropics: Combining ecological and economic goals. In: Vandermeer, J.H. (ed.). Tropical agroecosystems. CRC Press,Ne w York. Perfecto, I., Mas,A. , Diestch, T. and Vandermeer, J. 2003. Conservation ofbiodiversit y in coffee agroecosystems: atr i taxa comparison in southern Mexico. Biodiversity and Conservation 12: 1239-1252. Perfecto, I., Vandermeer, J.H., Lopez-Bautista, G., Ibarra-Nunez, G., Greenberg, R., Bichier, P. and Langridge, S. 2004. Greater predation in shaded coffee faros: the role of resident Neotropical birds. Ecology 85(10): 2677-2681. Petit, L.J., Petit, D.R., Christian, D.G. and Powell, H.D.W. 1999. Bird communities of natural and modified habitats inpanama . Ecography 22:292-304 .
32 Pimentel, D., Stachow, U., Takacs, D.A. and Brubaker, H.W. 1992. Conserving biological diversity in agricultural/forestry systems. Bioscience 42: 354-362. Pineda, E., Moreno, C, Escobar, F. and Halffte, G. 2005. Frog, bat, and dung beetle diversity in the cloud forest and coffee agroecosystems of Veracruz, Mexico. Conservation Biology 19:400-410 . Ramamoorthy, T.P., Bye, R., Lot, A. and Fa, J. (ed.). 1993. Biological biodiversity of Mexico: Origins and distribution. Oxford University Press,Ne w York. Rendon-Rojas, M.G. 1994.Estudio de laHerpetofaun a de la Zona Cafetalera de Santiago Jalahui, Oaxaca, B.Sc.thesis , EscuelaNaciona l de Ciencias Biologicas, Instituto Politecnico,Nacional , Mexico City. Rice, R. and Ward, J. 1996. Coffee, Conservation and Commerce in the Western Hemisphere: How Individuals and Institutions Can Promote Ecologically Sound Farming and Forest Management inNorther n Latin America, Smithsonian Migratory Bird Center andNatura l Resources Defense Council, Washington, D.C. Robbins, C.S. Dowell, B.A. , Dawson, D. K., Colon, J. A., Estrada, R., Sutton, A., Sutton, R. and Weyer, D. 1992.Compariso n ofNeotropica l migrant landbird populations wintering in tropical forests, isolated forest fragments, and agricultural habitats. In: Ecology and Conservation ofNeotropica l Migrant Landbirds, Hagan, J.M. and Johnson, D.W., Eds., Smithsonian Institution Press, Washington, D.C. Rosenzweig, M.L.1999. Species diversity. In: McGlade, J. (ed.).Advance d theoretical ecology: principles and applications. Blackwell Science, Oxford, UK. Rosenzweig, M.L. 2001. Loss of speciation rate will impoverish future diversity. PNAS 98: 5404-5410. Rosenzweig, M.L. 2005.Avoidin g mass extinction: Basic and applied challenges.24 th Annual Midwest Ecology and evolution Conference. Dept. of ecology and Evolutionary Biology, University ofArizona , Tucson. Sadeghian, S. 2000. Diversidad de mesoorganismos del suelo en lazon a cafetera, paper presented in Memorias Foro Internacional Cafe y Biodiversidad, Federation Nacional de Cafeteros de Colombia, Ed., Chinchinil, Colombia.
33 SAGARPA. 2005.Avanc e de siembras y cosechas perennes 2005. [Electronic version]. Retrieved January 2, 2006, from http://www.siap.sagarpa.gob.mx/ar_comagr2c.html Stephens, S.E., Koons, D.N., Rotella, J.J. and Willey, D.W. 2003. Effects of habitat fragmentation on avian nesting success: a review of the evidence at multiple spatial scales. Biological Conservation 115: 101-110. Stork,N.E . and Brendell, M.J.D. 1990. Variation in the insect fauna of Sulawesi trees with season, altitude and forest type. In: Knight, W.J. and Holloway, J.D. (eds.). Insects and the Rain Forests of South East Asia (Wallacea). Royal Entomological Society, London, UK. Taylor, P.D. and Merrian, G. 1995. Habitat fragmentation and parasitism of a forest damselfly. Landscape Ecology 11: 181-189. Tejeda-Cruz, C. and Sutherland, W.J. 2004. Bird responses to shade coffee production. Animal Conservation 7: 169-179. Terborgh, J. and Weske,J.S . 1969. Colonization of secondary habitats by Peruvian birds. Ecology 50:765-782 . Thiollay, J.M. 1995.Th e role of traditional agroforests inth e conservation of rain forest bird diversity in Sumatra. Coservation Biology 9:335-353 . Toledo, V.M. 2005.Repensa r la conservation: ^areas naturales protegidas o estrategia bioregional? Gaceta Ecologica 77:67-82 . Torres,J .A . 1984.Diversit y and distribution of ant communities in Puerto Rico. Biotropica 16:296-303 . Tosepan Titataniske. 2005.Agricultur a sustentable. [Electronic version]. Retrieved January 2, 2006, from http://www.tosepan.com/index.htm Trzcinski, K.K., Fahrig, L. and Merriam, G. 1999. Independent effects of forest cover and fragmentation on the distribution of forest breeding birds. Ecological Applications 9:586-151. Van de Voort, M. and Greenberg, R. 1999. Por que a las aves migratorias les fascina el cafe? El Jarocho Verde 11: 52-54.
34 Vennini, J.P. 1994.Nearti c avian migrants in coffee plantations and forest fragments of southwestern Guatemala, Birds Conservation International 4: 209-232. Villard, M.A., Merriam, G. and Maurer, B.A. 1995.Dynamic s in subdivided populations of Neotropical migratory birds in atemperat e fragmented forest. Ecology 76:27-40 . Wiegand, T., Revilla, E. and Moloney, K.A. 2005.Effect s ofhabita t loss and fragmentation onpopulatio n dynamics. Conserv. Biol.19(1) : 108-121. Wilson, D.S. 1992.Comple x interactions in metacommunities, with implications for biodiversity and higher levels of selection. Ecology 73: 1984-2000. Winker, K. 1998.Recen t geographic trends inNeotropica l avian research. Condor 100(4): 764-768. With, K. A. and Crist, T.O. 1995.Critica l thresholds in species' response to landscape structure. Ecology 76:2446-2459 . With, K.A. and King,A.W . 1999.Dispersa l success on fractal landscapes: a consequence of lacunarity thresholds. Landscape Ecology 14:73-82 . Wunderle J. 1999.Avia n distribution in Dominican shade coffee plantations: area and habitat relationships. Journal of Field Ornithology 70: 58-70. Wunderle,J . and Latta, S. 1994.Overwinter turnover ofNearti c migrantswinterin g in small coffee plantations in Dominican Republic, Journal of Ornithology 13: 54-77. Wunderle, J. and Latta, S. 1996.Avia n abundance in sun and shade coffee plantations and remnant pine forest inth e Cordillera Central, Dominican Republic. Ornitologia Neotropical 7: 19-34. Wunderle, J. M. and Latta, S. C. 1998.Avia n resource use in Dominican shade coffee plantations. Wilson Bulletin 110: 271-281. Wunderle, J. and Wide, R.B. 1993.Distributio n of overwintering Neartic migrants in the Bahamas and Greater Antilles. Condor 95:904-933 .
35
Chapter 2
Monitoring Neotropical birds Efficiency of amodifie d double-observer point counts approach versus mist-netting
Euridice Leyequien, Willem F. deBoer, GarryBakker, Reinoud Vermoolen, Samuel Lopez de Aquino Abstract
Monitoring bird community trends is central in conservation biology, providing information of species diversity and their status. However, the use of different sampling techniques for collecting information may result in disparities in assessing species richness.Poin t counts and mist netting are frequently used for assessing bird species richness. Point counts and mist netting have been compared inpreviou s studies;howeve r the emphasis has been on breeding orwinterin g grounds whereas little on migration. Bird monitoring during migration could produce valuable information about long-term community changes.Assessin g relative biases,cost s and efficiency ofbot h techniques would enable the optimisation inth e design of large-scale monitoring. Moreover, Neotropical avifauna with a high diversity isworldwid e recognised asthreatene d by habitat loss and fragmentation. In our study we compare the efficiency, limitations, and biases of a modified double-observer point count approach versus mist netting in estimating species richness and detecting understory and migrant species. We evaluated cost-efficiency aspects to optimise sampling protocols.W e found that the point count technique was the most effective intota l species richness completeness and presented lower total effort incompariso n to mist netting. Point counts outperformed mist netting in the detection of new bird species, even after a large sampling effort. Mist netting significantly detected a higher proportion ofunderstor y species in comparison to point counts,thoug h we found opposite results for migrant species. Finally, the cost-efficiency analysis showed that the modified double-observer point counts required lesstota l effort thus decreasing total monetary costs compared to mist netting.
38 Introduction
Monitoring community trends is central in conservation biology, aimed at providing reliable information of species diversity and their status, for fast and efficient identification of conservation priorities. Quantifying the species richness ofbir d communities has a crucial role in ecological research (Huston 1994,Rosenzwei g 1995), conservation planning (Bibby etal. 1992, Stotz etal. 1996),an d environmental impact assessments (e.g.; Fjeldsa 1999). Species lists are a basis for conservation efforts to identify areas with high species richness or high level of endemism (Peterson etal. 1998,Myers et al.2000 , Rojas-Soto et al. 2002). However, the use of different techniques to obtain this information mayresul t in considerable disparities in assessing species richness (Thomas 1996, Poulin etal. 2000 , Godoy-Bergallo etal. 2003) , and therefore may result in strong biases in describing local and regional species patterns. Moreover, the economic costs associated with the different sampling techniques for assessing species richness represent an important factor influencing the final assessment.
Neotropical avifauna comprises a high diversity and isworldwid e recognised as threatened by habitat loss and fragmentation due to human activities (Winker 1998). Despite the enormous effort of ornithologists and natural resource managers in counting terrestrial birds, knowledge about Neotropical birds is scanty compared for example with Neartic birds,whic h is in itself a major concern if long-term conservation goals are to be accomplished. However, this data iscrucia l in order to quantify species' distribution, bird- habitat relationships, orpopulatio n and community trends. Point counts and mist net techniques are some ofth e most frequently used for assessing bird species richness. However, both techniques have inherent biases for bird detection (e.g.;neglectin g secretive species,thos e with infrequent vocalisations, such as understory or high canopy species) (Bibby etal. 1992)whic h complicate a reliable comparison between e.g.; local and regional scale. Despite a number of studies comparing point counts and mist netting techniques, most ofth e emphasis has been centred on breeding or wintering grounds (Rappole et al. 1998, Ralph and Fancy 1995)wherea s little attention was paid during migration (Wang and
39 Finch 2002). Bird monitoring during migration could produce invaluable information about the long-term community changes ofresident and migratory Neotropical birds, improving effective conservation-oriented research and management. In addition, a cost-benefit analysis for the implementing of these techniques can assist in providing guidelines for cost effective monitoring and research. In thispape r we compare the efficiency, limitations, and biases of a modified double-observer point count approach versus mist netting in estimating species richness during the period of migration. We also evaluated some economic aspects of either using onetechniqu e or a combination ofboth . These analyses allow us to optimise sampling protocols for monitoring Neotropical bird communities and estimate species richness.Eve nthoug h our analyses were mainly exploratory, we made anumbe r of a priori predictions based upon earlier findings: (1)poin t counts yields higher total species richness at lower sampling effort; (2) detectability ofunderstor y and migrant species ishighe r using mist nets.
Materials and Methods
Study area
The avian communities were sampled in theNort h Eastern mountain range of Puebla, within the Cuetzalan region, Mexico, which covers an area of~ 151,368ha . The North Eastern mountain ranges of Puebla is situated on the east migratory paths ofNeartic - Neotropical birds that winter or refuel inthi s region. The region ispredominantl y a traditional shade coffee agro-ecological zone, which covers an altitudinal range between 300 to 1200 m. The major agricultural system istraditiona l polyculture (traditional shade coffee plantation),whil e other land covers, such as seasonal crops (i.e.;maiz e and beans) and grazing lands are also present inth e area. Forested areas aretypicall y remnants of tropical evergreen forest with low montane rain forest in the higher areas (between 7000 and 2500 m) and lowland rain forest in the lower areas (200-700 m).Th e mean annual rainfall is 3,500 mm without a defined dry season; and the mean annual temperature is22 - 26°C.
40 Bird survey
We performed abir d survey during the migratory season from September to November 2003 (Leyequien etal. submitted). The period comprised sequential rounds, visiting six different locations within the overall area, which were selected using a stratified sampling technique, that minimised heterogeneity among plots (~ 1 ha).Al l bird surveys were conducted within traditional coffee plantations. Because all birds were sampled in a single habitat type (traditional coffee plantations), variation in species composition among different plots is assumed to represent natural fluctuations and isno t created by habitat differences.
Point counts. We used amodifie d double-observer approach (Nichols etal. 2000) , i.e.; at each single point count two observers were registering all visual and acoustic detections, each covering a range of 180° (atota l of 360°)wit h a fixed-radius of 25 m. in ate n minute period for each single count (Dawson et al. 1997, Vielliard 2000). Communication between observers was performed to avoid double record, and individual field notes were recorded. In addition, to avoid counting bias in the sampling, the order ofth e point counts was alternated in every visit. A distance of 100m betwee n point counts was used to avoid dependency between points' data. A number of 10-12poin t counts were carried out in each location per day. The total effective sampling effort per plot per day was ~12hours .
Mist net. We used five 12-mmis t nets (2.5 X 12m ) with 30 mm mesh in each plot. As a standard, five mist nets (60 m) were used at each location. Nets were open from approximately sunrise till noon. The nets within aplo t were placed using arando m setting. The total effective sampling effort perplo twa s 170.5hours .
41 Statistical analysis
Species accumulation curves andSmax. The species accumulation curves for point counts and mist nets were calculated using apredictiv e approach bytakin g the number of survey days as sampling effort. To eliminate the influence of the order in which days were added toth e total,th e sample orderwa s randomised 1000time s using Estimates software (Colwell 2004). This produces smoothed accumulation curves by repeated random reordering of the samples (Longino and Colwell 1997). Thereafter, these accumulation curves were used to fit predicted curves derived from the species collection of each method. The Pearsons correlation squared between the fitted curve and species accumulation curve was calculated. Thevalues of the fitted curve were used topredic t theS max for each method, using an exponential function, with halt after 100trie s with < 0.0001.
Totalspecies richness. In ordert o evaluate both sampling methods we established an estimate of actual richness of bird species occurring inth e overall area, which could provide a standard for comparison ofth e two sampling methods (as recommended by Verner 1985,Watso n 2004).W e used predictive equations to estimate the total richness of the overall area, i.e.;th etheoretica l value for the maximum number of speciestha t exist (Smax), based on a finite number of samples (Colwell and Coddington 1994,Herzo g etal.
2002). TheS max was calculated using an exponential function with halt after 100trie s with < 0.0001; the total number of species recorded for the area (i.e.;th e species collection derived from both methods) was used for further comparison.
Species richnesspercentage of completeness. Theperformanc e of the methods was estimated by dividing the species richness estimates (Smaxderive d from the two methods) by thetota l richness (Smax) ofth e overall area, resulting in percentages of completeness and associated coefficients of variation.
42 Predicted effort andsampling optimisation. We determined the effort (i.e.;numbe r of
sample days) required to obtain the totalpredicte d species richness (Smax)usin g the software Effort Predictor V1.0 (Colwell 2004). In addition, in order to find an optimisation of the bird sampling, we calculated the added value ofth e sampling techniques. We compared the added value of mist netting days versus point counts days, calculating the effect of swapping a certain number ofpoin t counts sampling days against the same number of mist net samples,b y calculating total species richness from 100rando m combinations for each extra mist net sample.
Understoryand migrant species. The proportion ofunderstory species (including migrants and residents) and migrants detected by each technique were calculated. Shared and unique understory and migrant species for both techniques were derived. To test ifther e was a significant difference ofpercentag e (arcsine transformed) ofunderstory species between point counts and mist netting an analysis ofvarianc e (ANOVA)wa s performed. The same procedure was carried out for the migrant species.
Cost-efficiency analysis. The costs ofth e monitoring schemes were calculated using indicative recurrent monitoring costs, i.e.;USD/tota l sampling effort to detect 90%o f the total species richness. Costs by day arebased on locally-based monitoring by professional staff for the year 2003. Other day-costs such as field material and transportation were included inth e analyses. In addition, thepredicte d effort to obtain the total species richness (Smax) was included in the comparison.
Thebenefit s of using each technique were based on the total number of species,th e number of unique species,proportio n of understory species,numbe r of unique understory species, proportion of migrant species, andnumbe r of unique migrant species thatwer e detected by each technique.
43 Results
Thetota l species richness (Smax) estimated for the area is 136.7(k : 0.08, Pearsons
Correlation squared: 0.982), while the estimated species richness (Smax)fo r thepoin t count was 108.1 (k: 0.09, Pearsons Correlation squared: 0.994) and for the mist net was 70.1 (k: 0.03, Pearsons Correlation squared: 0.987) (Table 1).
Table 1. Cost-efficiency analysis of monitoring schemes: (a)point counts, (b)mist nets.
Monitoring scheme Point counts Mist nets Number of sampling days 23 23 Number of detected birds 2555 354
Predicted species richness (Smax) 108.1 70.1 Recorded species richness 115 79
Indicative recurrent costs of monitoring (USD/day to -„„ ^907
detect 90%o fS max) 26 71 Sampling effort in days (detecting 90%o fS max) 59 23 Unique species Understory species (%) including double counting 63 90 (N=30) Unique understory species 3 11
Migrant species (%) including double counting 75 61 (N=5l)
Unique migrant species 20 13
Species richnesspercentage of completeness. The comparison of the different methods using the percentage of completeness showed that the point count technique yielded the most complete species richness estimates with the highest consistency (83%, V= 3.3,n = 23), whereas the mist net technique had a lower completeness (53%, V= 4.2 , n =23) .
44 Predicted effort andsampling optimisation. We found that for point counts,th e sampling effort required to detect 90%o fth e total species richness required 26 sample days, whereas mist netting required 71 sample days (Figure 1).
/"(•>
Figure 1.Predicted sampling effortfor detecting 90% of the calculated totalspecies richness (Smax130.7) for (a)point count and (b)mist net techniques.
In addition, mist nets samples did not show a significant added value compared to point counts when point counts were swapped by mist nets (Figure 2).Th e analysis of variance showed that the percentage of understory species in the two survey techniques was significantly different (F1>44 = 19, P < 0.001) (Figure 3). The number of understory species detected by the point counts was 19(6 3 %), whereas 27 (90%) understory species were detected using mist netting. From the 19understor y species recorded by thepoin t counts only 3wer e unique to this technique, whereas from the 27 understory species detected by mist nets, 11wer e unique.
45 130
g> o a> 120 ri o. U) XI J) 5* EQ.
100 10 11 Mistnettin gsamplin gday s
Figure 2.Added value, i.e.number ofdetected species, ofmist net sampling days when substituting a certain number ofpoint countssampling days bymist netsamples (e.g.1-10 mistnet sampling days as shown in the bar chart). Error barsshow 95.0% CIof Mean. Understoryand migrant species.
?3 1 2 Survey technique Figure 3. Significant differences in themean number of understory speciesfor the two survey techniques. 1 = Point counts, 2 =Mist netting. 46 Moreover, for migrant species,th e analysis demonstrated that the percentage of migrants was significantly different between the two survey techniques (Flj44 = 33,P <0.001 ) (Figure 4).Th e number of migrant species detected by the point counts was 38 (75 %), whereas 31 (61%)migran t species by mist netting. From the 38 migrant species recorded by the point counts 20wer e unique to this technique, whereas from the 31migran t species detected by mist nets, only 13wer e unique. - 1 10°i •.-i"/i. 0) en <2 50- Survey technique Figure 4.Significant differences in the mean number of migrantspecies for the twosurvey techniques. 1 = Point counts, 2 =Mist netting. Cost-efficiency analysis. The cost for detecting 90%o f the total species richness was 1207 USD/total sampling effort using point counts,wherea s with the use of mist netting the total costs for the samepercentag e would increase to 3297 USD/total sampling effort. Mist netting required an additional effort of4 5 sampling days than the point counts,wit h an additional cost of 2090 USD/total sampling effort. We found that 59 species were detected only by the point counts,wherea s 23 species were reported only from mist netting, both methods shared 56 species. Point count sampling detected ahighe r number ofuniqu e species,whil e detecting a lower proportion of 47 understory species and unique number ofunderstory species in comparison to mist netting. In addition, point counts sampling detected ahighe r proportion of migrant species and number ofuniqu e species ofmigrant s (Table 2). Table2. Totalnumber ofspecies detected bypoint counts and mistnetting. Understoryand migrant species. Order Family Species Understory1 Status PC MN Ciconiiformes Ardeidae Bubulcus ibis 0 0 X Cathartidae Cathartesaura 0 0 X Falconiformes Accipitridae Accipiter striatus 0 1 X Buteo platypterus 0 1 X Columbidae Columba flavirostris 0 0 X Columbinainca 1 0 X Columbinatalpacoti 0 0 X Leptotila verreauxi 1 0 X Zenaida asiatica 0 0 X Psittaciformes Psittacidae Amazona autumnalis 0 0 X Pionus senilis 0 0 X Cuculiformes Cuculidae Crotophaga sulcirostris 0 0 X X Piaya cayana 0 0 X Strigiformes Strigidae Glaucidium brasilianum 0 0 X X Apodiformes Apodidae Chaeturavauxi 0 0 X Trochilidae Amazilia Candida 0 0 X X Amazilia cyanocephala 0 0 X Amazilia yucatanensis 0 0 X X Archilochus colubris 0 1 X Campylopteruscurvipennis 0 0 X X Eugenes fulgens 0 0 X Hylochoris leucotis 0 0 X Lampornis amethystinus 0 0 X Selasphorus platycercus 0 1 X Trogoniformes Trogonidae Trogonelegans 0 0 X Coraciiformes Momotidae Momotusmomota 0 0 X Piciformes Picidae Dryocopus lineatus 0 0 X Melanerpes aurifrons 0 0 X X Melanerpes formicivorus 0 0 X Piculus aeruginosus 0 0 X X Piculus rubiginosus 0 0 X Sphyrapicus varius 0 1 X Passeriformes Dendrocolaptidae Lepidocolaptes affinis 0 0 X Sittasomus griseicapillus 0 0 X X Tyrannidae Camptostoma imberbe 0 0 X Contopus borealis 0 1 X Contopus cooperi 0 1 X Contopus perlinax 0 0 X Contopus sordidulus 0 1 X X Contopus virens 0 1 X Empidonax alnorum 0 1 X Empidonax flaviventris 1 1 X X Empidonax hammondii 0 1 X 48 Order Family Species Understory' Status PC MN Empidonax occidentalis 1 0 X X Empidonax traillii 0 1 X Empidonax wrightii 0 1 X Megarynchus pitangua 0 0 X X Mionectes oleagineus 1 0 X Mitrephanes phaeocercus 0 0 X X Myiarchus crinitus 0 1 X X Myiarchus tuberculifer 0 0 X X Myiarchus tyrannulus 0 0 X Myiodynastes luteiventris 0 0 X Myiozetetes similis 0 0 X X Pitangus sulphuratus 0 0 X Tityrasemifasciata 0 0 X Tyrannus couchii 0 0 X Tyrannus melancholicus 0 0 X Hirundinidae Streptoprocne zonaris 0 0 X Corvidae Cyanocoraxmorio 0 0 X Cyanocorax yncas 0 0 X X Troglodytidae Campylorhynchus zonatus 0 0 X X Henicorhina leucophrys 1 0 X X Henicorhina leucosticta 1 0 X X Thryothorus maculipectus 1 0 X X Troglodytes aedon 1 1 X X Troglodytesbrunneicollis 0 0 X Sylviidae Polioptila caerulea 0 1 X X Regulidae Regulus calendula 0 1 X X Turdidae Catharus mexicanus 1 0 X Catharus ustutatus 0 1 X Chatarus aurantiirostris 1 0 X Hylocichla mustelina 0 1 X Myadestes occidentalis 0 0 X Sialia sialis 0 0 X Turdusassimilis 0 0 X X Turdusgrayi 0 0 X X Mimidae Dumetella carolinensis 1 1 X X Melanotis caerulescens 1 0 X X Vireonidae Cyclarhis gujanensis 0 0 X Vireo gilvus 0 1 X Vireo griseus 0 1 X X Vireohuttoni 0 0 X Vireoleucophrys 0 0 X X Vireoolivaceus 0 1 X X Vireophiladelphicus 0 1 X Vireo solitarius 0 1 X X Parulidae Basileuterus belli 1 0 X Basileuterus culicivorus 1 0 X X Basileuterus rufifrons 1 0 X X Dendroica coronata 0 1 X Dendroica occidentalis 0 1 X Dendroica magnoliae 0 1 X Dendroica townsendi 0 1 X Dendroica virens 0 1 X X Geothlypis trichas 0 1 X Helmitheros vermivorus 1 1 X Icteria virens 1 1 X Mniotilta varia 0 1 X X 49 Order Family Species Understory1 Status PC MN Myioborus miniatus 0 0 X Oporornisformosus 1 1 X Parula americana 0 1 X Parula pitiayumi 0 0 X Seiurus aurocapillus 1 0 X X Seiurusmotacilla 1 1 X Setophaga ruticilla 0 1 X Vermivora celata 0 1 X Vermivorapinus 0 1 X X Vermivoraruflcapilla 0 1 X X Wilsoniacanadensis 0 1 X X Wilsoniacitrina 1 1 X X Wilsoniapus ilia 0 1 X X Chlorospingus 0 Thraupidae ophthalmicus 0 X X Cyanerpes cyaneus 0 0 X X Euphonia hirundinacea 0 0 X X Piranga leucoptera 0 0 X Piranga rubra 0 1 X Thraupis abbas 0 0 X X Cardinalidae Cyanocompsa parellina 1 0 X Passerina ciris 1 1 X Passerina cyanea 1 1 X Pheucticus ludovicianus 0 0 X X Saltator atriceps 0 0 X X Emberizidae Arremonops rufivirgatus 1 0 X X Atlapetes brunneinucha 1 0 X X Melospiza lincolnii 1 1 X Sporophila torqueola 1 0 X X Tiaris olivacea 1 0 X X Volatiniajacarina 0 0 X Icteridae Dives dives 0 0 X X Icterus galbula 0 1 X X Icterus graduacauda 0 0 X X Icterus gularis 0 0 X X Icterus spurius 0 1 X Psarocolius montezuma 0 0 X Quiscalusmexicamis 0 0 X Fringillidae Carduelispsaltria 0 0 X Coccothraustesabeillei 0 0 X Discussion In accordance with our first prediction, the point count was clearly the most effective survey technique interm s oftota l species richness completeness. Thepoin t count technique was the most consistent of both techniques showing the smallest internal variation (V= 3.3). Whitman etal. (1997 ) found similar results in which the species detection rate was higher for point counts (60%) than for mist netting (25%). Similarly, Wang and Finch (2002) 50 found that point counts detected 82%o f the species, whereas mist netting only 74%. Contrasting results showed that mist netting was more effective for detecting species (92%) than point counts (68%) (Rappole etal. 1998), or as twice as many species aspoin t counts detected (Gram and Faaborg 1997).Difference s between the techniques could be habitat and species dependent. Variation in the vertical structure ofvegetatio n could affect the capture rate of specific species (e.g.; canopy species). However, for some species it has been found that detectability could not be attributed only to vegetation structure but also by species behaviour (Wang and Finch 2002). Species behaviour plays an important role in the detection performance of both techniques, for example the food searching behaviour of Cathartesaura or the complete aerial activities ofStreptoprocne zonaris, makes them easy to detect but difficult to capture regardless ofhabita t type. In addition, our results showed thatpoin t count technique requires a lower total effort in comparison to the mist netting. In our study the sampling effort required by the point counts to detect 90%o fth e total species richness was 2.7 times lower than the required effort by the mist net technique, giving a higher overall efficiency. Identifying differences in sampling effort is critical in the assessments of species richness and composition; omitting these differences could produce erroneous assessments of diversity. Our second prediction was partially supported also by our results;w e found that even thoughth e overall efficiency of the mist netting is lower, this technique significantly detected ahighe r proportion of understory species in comparison topoin t counts. The number of unique understory species detected by mist netting was almost three times higher than detected by point counts. For example, secretive species, such as Catharus mexicanus, Catharusaurantiirostris, Mionectes oleagineus, Oporornisformosus andSeiurus motacilla, were only detected by mist netting. For specific cases like the migrant Seiurus motacilla (with low overall density), that are difficult to monitor accurately in its breeding and wintering grounds, mist netting isrelevant . It hasbee n suggested by Remsen and Good (1996)tha t any bird survey must be accompanied by intensive mist net sampling to avoid under representation of secretive and understory species. However, the use of mist nets 51 alone must be considered with caution; other studies showed that mist nets areno t a replacement for overall community sampling, detecting only 41% (57%i n our study) of the total species richness recorded in the area (Bierregaard 1990). In addition,pro s and cons of mist netting should be taken into consideration for successful monitoring, i.e.; the advantage ofmis t netting inreducin g biases created by variability in observer experience and qualification, should be compared tobiase s created by net avoidance, habitat structure and behavioural differences between species (Herzog etal. 2002) . However, for migrant species, point counts significantly detected more species than mist netting. Point counts detected ahighe r number of migrant species and more unique species. In the case of the species detected only bypoin t counts,th e failure of mist netting in detecting the species could be partly explained by its inherent bias to miss canopy species such as raptors, e.g.;Accipiter striatus orButeoplatypterus. Other notable species detected only by point counts areEmpidonax species (i.e.;Empidonax alnorum, Empidonax flaviventris, Empidonax traillii,Empidonax wrightii) and the majority of warbler species (e.g.;Dendroica coronata,Dendroica occidentalis, Dendroica townsendi, Dendroica virens,Icteria virens, Mniotilta varia). Rappole etal. (1998 ) found that point counts performed better than mist netting for detecting migrating Wilsoniapusilla warbler across habitats inMexico .However , other migrants that forage inth e understory were detected only by mist netting (e.g.;Passerina ciris,Passerina cyanea, Melospiza lincolnii). It has been suggested that during migration mist netting is a more suitable technique because many migrants are actively foraging during stopovers and therefore easier to catch, while they are less vocal and territorial (Wang and Finch 2002). However, we did not find evidence in our study that supports this suggestion. Finally, the cost-efficiency analysis showed that point counts require less total labour and therefore decreases monetary costs compared to mist netting. Theperformanc e of point counts in terms of species detections including migrants is greater than mist netting. Despite the evidentprofit s derived from point counts,th e use of mist netting should not be disregarded when designing amonitorin g programme asthi s technique showed a higher 52 detection ofunderstory and secretive birds.Fo r a successful design of monitoring, the relative costs and biases of both techniques must be considered. Besides, whenever time and resources are available, point counts andmis t netting should both bepar t ofth e monitoring scheme (Rappole etal. 1998). Acknowledgments We thank Dr. Adolfo Navarro Siguensa and Laura E. Villasenor Gomez for providing the mist nets, and Gabriela Garcia Deras for field assistance. This research was supported by the CONACYT grant 149557 to E. Leyequien. References Bibby, C.J., Burgess N.D. Hill DA. 1992.Bir d census techniques. Academic Press, California. Bierregaard, JR., O. 1990. Species composition and trophic organization ofth e understory bird community in a central Amazonian terra firme forest. In: Four Neotropical forests (A. H. Gentry, ed.)Yale Univ. Press,Ne w Haven, Connecticut. Colwell, R.K. 2004 .EstimateS : Statistical estimation of species richness and shared species from samples. Version 7. User's Guide and application published at: http://purl.oclc.org/estimates. Colwell, R.K. and Coddington, J.A. 1994. Estimating terrestrial biodiversity through extrapolation. Phil. Tran. Royal Soc.Lon. 345: 101-118. Dawson, D.K., Smith D.R., Robbins C.S. 1997.Poin t count length and detection of forest Neotropical migrant birds. In: Ralph C.J., Sauer J.R. and Droege S. (eds) Monitoring Bird populations by point counts. General technical Report PSW- GTR-149. Albany, California, pp 35-43. Fjeldsa, J. 1999. The impact of human forest disturbance on the endemic avifauna of the Udzungwa Mountains, Tanzania. Bird Conservation International 9:47-62. 53 Godoy-Bergallo, H., Esberard, C.E.L., Riberiro-Mello, M.A., Lins,V. , Mangolin,R. , Melo, G.G.S. and Baptista, M. 2003. Bat species richness in atlantic forest: What isth e minimum sampling effort? Biotropica 35 (2):278-288 . Gram, W.R. and Faaborg, J.1997. The distribution ofNeotropica l migrant birds wintering in the El Cielo Biosphere Reserve,Tamaulipas ,Mexico . Condor 99: 658-670. Herzog, S.K.,Kessler , M. and Cahill, T.M. 2002. Estimating species richness of tropical bird communities from rapid assessment data. Auk 119(3) : 749-769. Huston, M.A. 1994.Biologica l Diversity. CambridgeUniversity Press, Cambridge, United Kingdom. Longino,J . and Colwell, R.K. 1997.Biodiversit y assessment using structured inventory: capturing the ant fauna of atropica l rain forest. Ecol.Application s 7: 1263-1277. Myers,N. , Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. and Kent, J. 2000. Biodiversity hotspots for conservation priorities.Natur e 403: 853-858. Nichols, J.D., Hines, J.E., Sauer, J.R., fallon, F.W., Fallon, J.E. and Heglund, P.J. 2000.A double-observer approach for estimating detection probability and abundance from point counts.Auk 117(2) : 393^108. Peterson, A. T., A. G. Navarro & H. Benitez. 1998.Th e need for continued scientific collections: A geographic analysis of Mexican birds specimens. Ibis 140:288-294 . Poulin, B.,Lafebvre , G. and Pilard, P.2000 . Quantifying the breeding assemblage of reedbed passerines with mist-net and point-count surveys.J . Field Ornithol. 71(3) : 443^154. Ralph, C. J. and Fancy, S.G. 1995.Demograph y and movements of Apapane and Iiwi in Hawaii. Condor 97: 729-742. Rappole, J.H., Winker, K. and Powell, G.V.N. 1998. Migratory bird habitat use in southern Mexico: mist nets versus point counts.Journa l of Field Ornithology 69:635-643 . Remsen,J.V. , and Good, D.A. 1996.Misus e of data from mist-net captures to assess relative abundance inbir d populations. Auk 113: 381-398. Rojas-Soto, O., S. Lopez de Aquino, L. Sanchez-Gonzalez & B. Hernandez-Banos. 2002. La colecta cientifica en el Neotropico: el caso de las aves de Mexico. Ornitologia Neotropical 13:209-214 . 54 Rosenzweig, M.L. 1995.Specie s Diversity in Space and Time. Cambridge University Press, Cambridge, United Kingdom. Stotz,D.F. , Fitzpatrick, J.W., Parker III,T.A . and Moskovits, D.K. (eds.) 1996. Neotropical Birds: Ecology and Conservation. University of Chicago Press, Chicago. Thomas, L. 1996.Monitorin g long-term population change: Why arether e so many analysis methods? Ecology 1: 49-58. Verner, J. 1985.Assessmen t of counting techniques. Current Ornithology 2: 247-302. Vielliard, J.M. 2000.Bir d community as an indicator ofbiodiversity : results from quantitative surveys. Brazil.A . Acad. Bras. Ci. 72 (3):323-330 . Wang, Y. and Finch, D.M. 2002. Consistency ofmis t netting andpoin t counts in assessing landbird species richness andrelativ e abundance during migration. Condor 104: 59-72. Watson, D.M. 2004. Comparative evaluation of new approaches to survey birds. Wildlife Research 31: 1-11. Winker, K. 1998.Recen t geographic trends inNeotropica l avian research. Condor 100(4): 764-768. Whitman, A. A., Hagan Iii, J. M. and Brokaw,N.V.L . 1997.A comparison oftw o bird survey techniques used in a subtropical forest. Condor 99: 955-965. 55 If*1"**'- • * ' v. *1 *. 1, "\,'J1A',?;; Chapter 3 Fragmentation and habitat loss: species richness responses in an avian metacommunity in Mexico. Euridice Leyequien, WillemF. deBoer, VictorM. Toledo Abstract In this paper we argue that fragmentation and habitat loss is of central importance in understanding the effects of spatial structure on species richness. We consider that an approach which separates and quantifies the effects of fragmentation and habitat loss on species richness will be more adequate to understand the influence of each process inth e community structure. However, many studies dono t separate the individual relative effects of fragmentation and habitat loss, leading small room for generalisations on patterns of species richness influenced by fragmentation and habitat loss.W e examined the relative individual and combined influence of these two factors on species richness ina n avian metacommunity. The first hypotheses in our study was (1) species richness will show a hump-shaped relationship (unimodal response) as a result of habitat fragmentation at landscape level, (2) as the amount of habitat in the landscape declines, extinction thresholds will be exceeded, creating a linear negative effect on species richness. Moreover, difference in explanatory power of individual and combined influence of both factors was compared. The results indicate that the response of species richness to habitat fragmentation shows a unimodal response at landscape level, and anegativ e response tohabita t loss.Th e combined influence of fragmentation and habitat loss did not offer a better approximation of species richness response. This suggests that there isn o interaction between the effects of fragmentation and habitat loss.Du e to growing human perturbation inth e landscape, crucial decisions for conservation and landscape management have to be made. We argue that influencing fragmentation levels will mitigate the negative effects of habitat loss on species richness. 58 Introduction The role of spatial dynamics in understanding differences in species diversity, mainly through fragmentation and habitat loss,i s increasingly becoming an important subject in ecology (Fahrig 2003, Rosenzweig 1995).Th e metacommunity theory has recently emerged as acommo n framework for scientistst o understand the effects of fragmented systems in community ecology (Leibold etal. 2004) . In ametacomunity , a set of local communities is linked by dispersal of species (Gilpin and Hanski 1991, Wilson 1992). Landscape fragmentation andhabita t loss strongly influence these (meta)community patterns and dynamics, disrupting important ecological processes for species persistence (e.g.;mortality , re-colonisation, or reproductive rates) (Bender etal. 1998,Fahri g 1997, 2001, Fahrig and Merriam 1994, Stephens etal. 2003 , Villard etal. 1995,Wiegan d et al. 2005). Habitat fragmentation leadst o configuration changes inth e landscape, increasing the number and isolation ofpatches , and decreasing patch sizes. Changes in the spatial structure of landscapes through fragmentation or aggregation of natural habitats can alter the abundance of single species,th e species diversity patterns, and thereby the community assemblage (Johnson etal. 1992,Robinso n and Wilcove 1994).Additionally , when high fragmentation levels occur dispersal probabilities decrease, sotha t species with low dispersal capacities can no longer colonize these isolated patches, consequently leading toa parallel reduction of local diversity (Leibold etal. 2004) . Moreover, at low fragmentation levels species richness will decline as a dominant competitor will invade each local community as one large community is created (Mouquet and Loreau 2002).A t intermediate fragmentation levels, species arebelieve d to be able to colonize isolated patches, but fragmentation levels are still not so low that a dominant competitor can invade all patches and control local diversity. Therefore we would expect aunimoda l response of species richness inrespons e to an increase in fragmentation. 59 Habitat loss,describe d here asth e reduction of total habitat area, has a consistently negative effects on species diversity. The destruction of habitat affects, among others, breeding and dispersal success (Belisle etal. 2001 , Kurki etal. 2000 ,Wit h and King 1999), trophic chain length (Komonen etal. 2000 ) and species interactions (Taylor and Merrian 1995).A n increase in habitat loss gradually exceeds species-specific extinction thresholds (With and Crist 1995). Many studies, describing the relative effects of fragmentation and habitat loss,resulte d in varying or even conflicting conclusions. Generally, negative linear trends have been predicted by theoretical models aswel l as by field studies,howeve r alsopositiv e effects on species richness have been found in landscapes with a large number of smaller patches. The dissimilarities among these conclusions are partially duet o differences in defining habitat fragmentation, where many researches do not separate the effects of habitat loss from the configuration effects of fragmentation (Fahrig 2003) producing unclear results leaving small basis for generalisations. An approach which quantifies both the effects ofth e degree offragmentatio n and habitat loss on species richness will be more adequate. We argue that the impact ofhabita t fragmentation on species richness in a metacommunity can not be adequately explained by linear trends;non-linea r relationships would perform better. We assume that species richness will decrease at high and low fragmentation levels because first athig h levels dispersal probabilities decrease leading to areductio n of local diversity, and at low levels a dominant competitor will invade each local community. In addition, we assume that at intermediate levels, species colonise isolated patches where a dominant competitor can invade all patches and control local diversity. Moreover, we expect a negative trend in species richness as aresul t ofhabita t loss,w e assume that an increase ofhabita t loss gradually exceeds species-specific extinction thresholds. The predictive power of both habitat fragmentation and habitat loss in explaining species richness at metacommunity level will be greater than the individual influence. 60 We examined the relative individual and the combined influence of fragmentation and habitat loss on species richness in an avian metacommunity in a Mexican coffee agro- ecological area. The aims of our study were to test the hypotheses that (1) species richness will show a hump-shaped relationship (unimodal response) as aresul t of habitat fragmentation at landscape level, and (2) asth e amount ofhabita t in the landscape declines, extinction thresholds will be exceeded, creating a linear negative effect on species richness. Moreover, the difference in explanatory power ofth e individual and combined influence of both habitat fragmentation and habitat loss will be compared, and we expect that the combined model performs significantly better than each of the individual models. Material and methods Study site The avian communities were sampled in the North Eastern mountain range of Puebla, within the Cuetzalan region, Mexico,whic h covers an areaof - 151,368ha . The region is predominantly atraditiona l shade coffee agro-ecological zone,whic h covers an altitudinal range between 300 to 1200 m. The major agricultural system istraditiona l polyculture (traditional shade coffee plantation), while other land covers, such as seasonal crops (i.e.; maize and beans) and grazing lands are also present in the area. Forested areas are typically remnants of tropical evergreen forest with low montane rain forest in the higher areas (between 700 and 2500 m) and lowland rain forest in the lower areas (200-700 m). The mean annual rainfall is 3,500 mm without a defined dry season; and the mean annual temperature is 22-26°C. We selected eight landscape frames of 10x 10k m for the study, for the characterisation of the metacommunity. All landscapes were selected using a stratified sampling design to minimise heterogeneity among the landscapes using a supervised classification image derived from an ETM satellite image (2003). All landscapes were selected within traditional shade coffee plantation areas where altitude, exposition and age of plantation were homogenised asmuc h as field conditions allowed. 61 Bird survey andspecies richness We performed a bird survey during the migratory and breeding seasons from November 2002 toNovembe r 2003 (Leyequien etal. submitted). Theperio d of sampling consisted of weekly sequential rounds visiting all of the eight landscapes in each round. For bird detection we used standard point count techniques (Bibby et al. 1992)wit h a modified double-observer approach, combining visual and acoustic bird detections (Parker 1991). Bird observations were recorded over 360° with a fixed-radius of2 5 m, and we used ate n minute period for each single count (Dawson etal. 1997,Vielliar d 2000).A distance of 100 mbetwee n point counts was used to avoid dependency between points' data. We restricted our bird species data sett o only year-round resident birds.Th e total number of registered species in all the landscapes consisted of 112residen t birds, representing 12orders , 28 families and 82gener a (Annex 1). Species richness estimators for the nine bird communities were calculated using the software Estimates V7.0 (Colwell 2004) and Effort Predictor VI.0. We estimated theS max richness estimator using an exponential model with 100iteration s after which improvements were < 0.0001;w e also estimated the first-order Jacknife richness estimator, using mean among runs (Burnham and Overton 1978,1979, Heltshe and Forrester 1983, Smith and van Belle 1984),an d the Chao 1richness estimator , also using mean among runs (Chao 1984), asrespons e variables. For the calculations we used the randomization protocol for estimators' prediction using sampling without replacement and thebias - corrected formula for Chao 1,wit h an upper abundance limit for rare or infrequent species of 10. Landscape fragmentation and habitat loss We used Fragstats 3.3 to compute the landscape metrics used to quantify landscape configuration (fragmentation) and habitat loss. The fragmentation explanatory variable set consisted of: contiguity index distribution for each land cover class (i.e.;coefficien t of 62 variation parameter), percentage of like-adjacencies involving the corresponding land cover class, landscape fractal index distribution, landscape shape index, landscape contagion index, landscape percentage of like adjacencies, connectance index, and landscape Shannon's diversity index. The habitat loss explanatory variable set consisted of total class area, percentage of landscape comprised of the corresponding class,an d number ofpatche s ofth e corresponding class. We used multivariate statistics to test for the effects of habitat fragmentation and habitat loss variables' sets on the avian species composition and relative abundances. We utilised a redundancy analysis (RDA) (Jongman et al. 1987)t o quantify the strength ofth e variables that best explained the variation inth e species data. We used a logtransformatio n of species data, and stepwise forward selection of explanatory variables with Monte Carlo permutation tests (9999 number ofunrestricte d permutations) where the residuals of the reduced model were permuted (thereduced-mode l method was selected to reduce a type I error). After selecting the canonical axis that best explained the variation in species data, we used the variance partitioning procedure (Borcard etal. 1992)t o quantify the marginal and conditional effects ofth e significant explanatory variables (a = 0.05) and the unexplained variation accounted for by either fragmentation orhabita t loss explanatory sets. Themodels To test the fragmentation-dispersal and habitat loss hypotheses, we used non-linear regression models for curve fitting, using GraphPad Prism V4.0.Fo r comparison of fit we used three different models as follows: Polynomial First Order (straight line), Polynomial Second Order and a Gaussian distribution. To discriminate among models we used the extra sum-of-squares F test for nested models with/? = 0.05 (Pesaran and Deaton 1978),o r the Akaike's Information Criterion (AIC) for non-nested models (Bozdogan 2000). 63 For testing our third hypothesis we developed a combined model using both fragmentation and habitat loss explanatory sets using a multiple non-linear regression (P= 0.05) with iterative estimation algorithms (SPSS V12.0.1.). TheF value was estimated using the sum of squares for the regression and the residual, and the/? value referred to the corresponding F-statistics (p = 0.05). Results From the eight selected landscape fragmentation variables,th e contiguity index for the class secondary vegetation (p =0.0060) , and the class forest (p =0.0200) , significantly explained 0.526 ofth e variance inth e species data according to the RDA. The marginal effect explained by the fragmentation explanatory variable was 23.7 % accounted for by the contiguity index for the class secondary vegetation, and 17.4% accounted for by the contiguity index for the class forest. Both variables explained 11.5 % ofvarianc e by their conditional effect, and ultimately the unexplained fraction of variance was 58.9 %. For the habitat loss explanatory variables,tota l area for the class secondary vegetation (p = 0.008) and the number ofpatche s for the class coffee plantations (p =0.05 ) explained 0.478 ofth e variance in species data. Moreover, the total area for the class secondary vegetation explained 15.4 % of the community composition by its marginal effect, whereas this was 25.9 % for the number ofpatche s for the class coffee plantations. Both variables explained 6.5 % of variance by their conditional effect, whereby theunexplaine d fraction of variance was 58.7 % (Fig. 1). 64 , *—-«—* Utwwohnd vamra 70 | 60 J 60 • »*l - I 40 Wrajnuie" ' 30 Ita-tfUlfi'twl wteW laaraaisine 20 MargEisteffec t secondary vegetation 10 Marginaleffec t j secondary vegetation Fragmentation Explanatoryvariables 'se t Figure 1. Thevariance partitioning of thespecies data, showing thepercentage of variance explained by the marginal effect of the variables contiguityfor the class secondary vegetation and the variable contiguityfor the classforest (fragmentation explanatory set); and the variables totalarea for theclass secondary vegetation and the variable number of patchesfor the class coffeeplantations (habitat lossexplanatory set). In addition, the conditional effectsfor the variables of bothfragmentation and habitat loss explanatory sets, and the unexplained variance aredepicted. Themodels For the fragmentation explanatory set,result s demonstrated that for the variable contiguity for the class secondary vegetation, the Gaussian distribution model had the besfitt i n explaining the species richness (Jacknife estimator). The AIC test showed that based on the comparison offits, th e Gaussian distribution model had the highest (62.05 %)probabilit y in describing the observed response patterns in species richness (p= O.0001; R2 = 0.1848), with aprobabilit y ratio of 1.7. For the second variable, contiguity for the forest class,base d on the AIC test, the Gaussian distribution model was also the most likely to be correct in explaining the species richness (Chao 1estimator) . Itha d a 54.8 %probabilit y of being correct,wit h aprobabilit y ratio of 1.5 (p =O.0001 ; R2 = 0.4888) (Table 1;Fig .2) . 65 (a) (b) 105-1 100< 95< *%-.--„ 90< 1- m BO. 75- 70. ' •» 65' — , t, . , 75 SO SS 90 95 100 105 100 105 110 115 120 125 Contiguity index (secondary vegetation) Contiguity index(forest ) 60< (c) (d) .--•* 50- 1 1500 2000 2500 3000 3500 WOO 2500 300D 3500 4000 4500 5000 Ciassare a (coffee plantations) Classare a (secondary vegetation) Figure 2. Predicted modelsfor fragmentation and habitat loss (CI95%) describing the observed response inbird species richness (closed dots). A humped-shaped relationship between fragmentation andspecies richness (contiguity indexfor secondary vegetation, P = <0.0001,R 2 = 0.1848;contiguity indexfor forest, P =<0.0001, R 2 =0.4888). A positive linear relationship between habitat lossand species richness (total class areafor coffee plantation, P =<0.0001, R 2 =0.4648; total class areafor secondary vegetation, (P= <0.0001, R =0.060). Habitat loss increases from left toright on the x-axis. Table 1. Comparison ofmodel fitting for fragmentation explanatory variables, best fit (*) wasselected based on theprobability to becorrect (%)and theR 2. *Probability of correctness. Fragmentation explanatory variables Explanatory variable Species richness Model AIC estimator Probability* (%) R2 /"-value Contiguity Index Jacknife Polynomial 2ndorde r 37.95 0.1810 (secondary vegetation) Gaussian 62.05* 0.1848 <0.0001 Contiguity Index Chaol Polynomial 2ndorde r 45.25 0.4866 (forest) Gaussian 54.8* 0.4888 <0.0001 66 As expected, for the habitat loss explanatory set, results showed that there is a negative linear relationship between the amount ofhabita t and species richness. For the explanatory variable class area for the class coffee plantation, we found that the first order polynomial had the best fit (p = <0.001; R2 = 0.4648) in explaining species richness (Smax estimator). Additionally, for the explanatory variable class area for the secondary vegetation, the first order polynomial also had the best fit. The extra sum-of-squares F-test showed that, based on the comparison of fits, the first order polynomial described the observed response patterns in species richness best (Chao 1estimator;/ ; = <0.0001,R 2 = 0.060; Table 2;Fig . 2). Table2. Comparison of modelfitting for habitat loss explanatory variables, models with P = 0.05were selected (*) that explained most variance of thespecies richness data (R2). S, Species richness. Habitat lossexplanator y variables Explanatory variable S estimator Model Extra sum-of-squares F test Hypothesis R2 P-value Total Class Area Smax Polynomial 1storde r Ho: accepted* 0-4648 <0.0001 (coffee plantation) Polynomial 2ndorde r H,: rejected 0.4064 st Total Class Area Chaol Polynomial 1 orde r Ho:accepted * 0060 (secondary vegetation) Polynomial 2nd order Hi rejected 0.5811 O.0001 The combined model that explained the species richness response best (Chao 1estimator ) had anR 2 =0.5 4 with/?= O.0001. TheAI Ctes t showed thatth e combined modelha s a low probability ofbein g correct in comparison to any ofth e individual models. The probability that the combined model was correct was 36.5 % against 63.5 % for all the individual models with aprobabilit y ratio of 1.8. 67 Discussion Firstly, our results indicate that the bird species richness show a unimodal response to habitat fragmentation at landscape level. This suggests that landscape configuration patterns (such as fragmentation levels) determine species richness, indicating that landscape patterns are central in understanding ecological processes (Bascompte and Rodriguez 2001, Turner 1989). In landscapes with intermediate levels of habitat fragmentation species richness was higher. Hence,ther e isa n optimum spatial structure at which species richness is highest, due to a balance between important ecological processes occurring at landscape scale, such as species competition and dispersal frequency. On the other hand, when one large community is created, at lowest fragmentation levels, aregionall y dominant competitor can invade each local community, reducing the local diversity. (Amarasekare andNisbe t 2001, Caswell 1978,Forbe s and Chase 2002, Mouquet and Loreau 2002, Taylor 1988).A t high fragmentation levels species richness in local communities will decrease and become homogenised, by gradually exceeding species-specific extinction thresholds (Leibold etal. 2004). Another determinant explaining this unimodal response isth e reduction of fragments isolation (i.e.;highe r connectedness) of suitable habitat (i.e.;forest , secondary vegetation) that increases the dispersal frequency. Consequently, dispersal-limited species and some rare species (which are typically driven to extinction quickly) can be maintained in local communities as transient fugitives orb y arescu e effect (Brown and Kodric-Brown 1977, Loreau and Mouquet 1999,Mouque t and Loreau 2002,Pullia m 1988,Shmid a and Wilson 1985). These results coincide with those obtained by other authors:highl y isolated habitat fragmentspresen t lower species richness, whereas fragments with higher connectedness had higher species richness (Baur and Erhardt 1995,va n Dorp and Opdam 1987. Furthermore, smaller patches with intermediate connectedness that areno t separated by an inhospitable matrix may support higher species richness.Additionally , in our study, the significant effect of connectedness ofth e forest and secondary vegetation classes could indicate that some forest specialists aswel l as generalists can coexist locally in these intermediate fragmented landscapes. 68 Secondly, the effect ofhabita t loss that we found was consistent with previous studies. Similar to our results, several studies found a negative effect of habitat loss on species richness (Collinge 1995,Fahrig 1997,Hol tet al. 1995a, 1995b,Olf f and Ritchie 2000, Robinson etal. 1992,Verboo m and Apeldoorn 1990). We found that as the amount of suitable habitat (i.e.;coffe e plantations, secondary vegetation) in the landscape declined, species richness responses showed anegativ e linear relationship. However, in the case of the variable area for the secondary vegetation class there is a lowpercentag e of variance explained. This might be explained by the fact that connectedness inthi s land cover class is more important than the area availability itself, and that the degree of habitat loss does not reach an extinction threshold (Andren 1994,Farin a 1997, Santos and Telleria 1997). Some studies found a threshold for habitat cover from which habitat fragmentation had no significant influence on species richness (Andren 1994,Fahri g 1997, McGarigal and McComb 1995). However, these latter findings considered forested areas, which may be significantly more accessible for bird dispersal than agricultural areas (Villard etal. 1999), as isth e case in our study in which fragments of forest were embedded in an agricultural matrix. A matrix with agricultural areas as predominant cover may influence not only the species composition but also the dispersal threshold of species. Interestingly, the combined influence ofbot h fragmentation and habitat loss did not offer abette r approximation of the species richness response ofth e bird metacommunity than the individual effect. This suggests that the effects of a single process (fragmentation orhabita t loss) do not show any interaction. We believe that by understanding the impact ofbot h fragmentation and habitat loss on species richness separately, generalisations can be made regarding species richness responses in metacommunity at landscape level. Conclusion Our results address fundamental issues in ecology, such asth e relationship between species richness and habitat fragmentation and habitat loss.Ou r univariate models atth e landscape scale explain significantly the species richness unimodal response to habitat fragmentation 69 and the negative linear response to habitat loss.Moreover , the conservation of species or assemblages in aregio n requires knowledge ofth e relevant indicators for conservation and the underlying structuring forces. We have shown that species richness responses to habitat fragmentation follow a hump-shaped pattern, and that there is a clear negative linear relationship between species richness and habitat loss,whic h obviously has major implications for conservation strategies. Fragmentation does not always have a negative effect on species richness,bu t there seems to be an optimal level of fragmentation at which species richness ishighest . In the context ofth e growing perturbation inth e landscape by human activities,crucia l decisions for conservation and landscape management will have to be made. It seemstha t by influencing fragmentation levels,th e negative effects of habitat loss could be mitigated. Acknowledgements Thanks to Dr. Ir. Andre Schaffers of the Nature Conservation and Plant Ecology Group and Drs. Nicole Heuermann of the Resource Ecology Group at Wageningen University for statistical consultation. We thank the "Society for the advancement of research in the tropics" for the grant bestowed to E. Leyequien. This research was supported by the CONACYT grant 149557 to E. Leyequien. References Andren, H. 1994. Effects of fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: areview . Oikos 71: 355-366. Amarasekare, P.an d Nisbet, R. 2001. Spatial heterogeneity, source-sink dynamics, and the local coexistence of competing species.America n Naturalist 158:572-584 . Bascompte, J. and Rodriguez, MA. 2001. Habitatpatchines s and plant species richness. Ecology Letters 4: 417-420. Baur, B. and Erhardt, A. 1995.Habita t fragmentation andhabita t alterations: principal threats to most animal and plant species. GAIA 4: 221-226. 70 Belisle, M., Desrochers, A. and Fortin, M.J. 2001. Influence of forest cover on the movements of forest bords: ahomin g experiment. Ecology 82: 1893-1904. Bender, D.J., Teschendorf, L. and Fahrig, L.1998. Evaluation of patch isolation metrics for predicting animal movement in binary landscapes. Landscape Ecology 15: 131- 143. Bibby, C.J., Burgess,N.D . and Hill, D.A. 1992. Bird census techniques, 1st edn. Academic Press, California. Borcard, D., Legendere, P. and Drapeau, P. 1992. Partialling out the spatial component of ecological variation. Ecology 73(3), 1045-1055. Bowman, J., Cappuccino,N . and Fahrig, L.2002. Patch size andpopulatio n density: the effect of immigration behavior. Conservation Ecology 6:9. Bozdogan, H. 2000. Akaike's Information Criterion and Recent Developments in Information Complexity. Journal of Mathematical Psychology 44(1),62-91(30) . Brown, J.H. and Kodric-Brown, A. 1977.Turnove r rates in insular biogeography: effect of immigration or extinction. Ecology 58:445-449 . Burnham, K.P. and Overton, E.W. 1978.Estimatio n ofth e size of a closed population when capture probabilities vary among animals. Biometrika 65:623-633 . Burnham, K.P. and Overton, E.W. 1979. Robust estimation ofpopulatio n size when capture probabilities vary among animals. Ecology 60:927-936 . Caswell, H. 1978. Predator mediated co-existence: anon-equilibriu m model. American Naturalistic: 127-154. Collinge, S.K. 1995. Spatial arrangements ofpatche s and corridors in the landscape: consequences for biological diversity and implications for landscape architecture. Ph.D. dissertation. Harvard University, Cambridge, Massachutes. Colwell, R.K. 2004 .Estimates : Statistical estimation of species richness and shared species from samples.Versio n 7. User's Guide and application published at: http://purl.oclc.org/estimates. Chao, A. 1984.Non-parametri c estimation ofth e number of classes in a population. Scandinavian Journal of Statistics 11: 265-270. 71 Dawson, D.K., Smith, D.R. and Robbins, C.S.. 1997. Point count length and detection of forest Neotropical migrant birds. Pp. 181.In : Ralph, C.J., Sauer, J.R. and Droege, S. (eds.) Monitoring Bird populations by point counts. General technical Report PSW-GTR-149.Albany , California. Fahrig, L. 1997. Relative effects ofhabita t loss and fragmentation on population extinction. Journal of Wildlife Management 61: 603-610. Fahrig, L.1998.Whe n does fragmentation ofbreeding habitat affect population survival? Ecological Modelling 105:273-292 . Fahrig, L. 2001.Ho w much habitat is enough? Biological Conservation. 100:65-74 . Fahrig, L. 2003. Effects ofhabita t fragmentation on biodiversity. Annu. Rev. Ecol. Evol. Syst. 34:487-515 . Fahrig, L. and Merriam, G. 1994. Conservation of fragmented populations. Conservation Biology 8:50-59 . Farina, A. 1997.Landscap e structure aNd breeding bird distribution in a sub-Mediterranean agro-ecosystem. Landscape Ecology 12:365-378 . Forbes, A.E. and Chase, J.M. 2002. Therol e of habitat connectivity and landscape geometry in experimental zooplankton metacommunities. Oikos 96:433-440 . Gilpin, M.E. and Hanski, LA. 1991. Metapopulation dynamics: Empirical and theoretical investigations. Academic press, London. Heltshe, J. and Forrester, N.E. 1983.Estimatin g species richness usingjacknif e procedure. Biometrics 39: 1-11. Holt, R.D., Debinski, D.M., Diffendorfer, Gaines, M., Martinko, E., Robinson, G., and Ward, G. 1995a. Perspectives from an experimental study of habitat fragmentation in an agroecosystem. Pp.147-175 in Glen, D.M., Greaves, M.P., and Anderson, H.M., editors. Ecology and integrated farming systems. Wiley, New York. Holt, R.D., Robinson, G.R. and Gaines, M.S. 1995b.Vegetatio n dynamics in an experimentally fragmented landscape. Ecology 76: 1610-1624. Hoopes, M.F., Holt, R.D., Shurin, J.B.,Law, R., Tilma, D., Loreau, M. and Gonzalez, A. 2004. The metacommunity concept: a framework for multi-scale community ecology. Ecology Letters 7: 601-613. 72 Johnson, A.R., Wiens, J.A., Milne,B.T . and Crist, T.O. 1992.Anima l movements and population dynamics in heterogeneous landscapes. Landscape Ecology 7:63-75 . Jongman, R.H.G., Ter Braak, C.J.F. and Van Tongeren, O.F.R. 1987.Dat a analysis in community and landscape ecology. Cambridge University Press, Cambridge. Keitt, T.H., D.L. Urban, and B.T. Milne. 1997.Detectin g critical scales in fragmented landscapes. Conservation Ecology (online) 1(1):4 . http://www.consecol.org/vol1 /is s1 /art4 / Komonen,A. , Penttilae, R., Lindgren, M. and Hanski, I. 2000. Forest fragmentation truncates a food chain based on an old-growth forest bracket fungus. Oikos 90: 119-126. Kurki, S., Nikula, A., Helle, P. and Linden, H.. 2000. Landscape fragmentation and forest composition effects on grouse breeding success in boreal forests. Ecology 81:1985 — 1997. Leibold, M.A., Holyoak, M., Mouquet,N. ,Amarasekare , P.,Chase ,J.M. , Hoopes, M.F.,Holt , R.D., Shurin, J.B., Law, R., Tilman, D., Loreao, M. and Gonzalez, A. 2004. The metacomunity concept: a fraMework for multiscale community ecology. Ecology Letters 7:601-613 . Leyequien, E., De Boer, W.F. and Skidmore,A.K . (submitted) Linking species- environment relationships and multiple spatial scales in community ecology. Loreau, M. and Mouquet, N. 1999. Immigration and the maintenance of local species diversity. American Naturalist 154:427-440 . Mcgarigal, K. and Mccomb, W.C. 1995. Relationships between landscape structure and breeding birds in the Oregon Coast Range. Ecol. Monogr. 65:235-260 . Mouquet, N. and Loreau, M. 2002. Coexistence in metacommunities: the regional similarity hypothesis. American Naturalist 159:420-426 . Olff, H. and Ritchie, M.E. 2000. Fragmented nature: Consequences for biodiversity.3 rd International Workshop on Sustainable Land Use Planning, Wageningen. Parker, T.A. 1991.O nth e useo ftap erecorder s in avifaunal surveys.AU K 108(2),443-444 . Pesaran, M. H. and Deaton, A. S. 1978. TestingNon-neste d Nonlinear Regression Models. Econometrica 46, 677-694. 73 Pulliam, H.R. 1988.Sources , sinks, andpopulatio n regulation. American Naturalist 132: 652-661. Robinson, G.R., Holt, R.D.,Gaines , M.S.,Hamburg , S.P., Johnson, M.L., Fitch, H.S. and Martinko, E.A. 1992.Divers e and contrasting effects of habitat fragmentation. Science 257:524-526 . Robinson, S.K. and Wilcove, D.S. 1994.Fores t fragmentation in the temperate zone and its effects on migratory song birds. Bird Conservation International 4:223-249 . Rosenzweig, M.L. 1995. Species diversity in space and time. Cambridge University Press, Cambridge, united Kingdom. Santos,T . and Telleria, J.L. 1997.Efecto s de lafragmentatio n sobre las aves insectivoras forestales de dos localidades europeas.Ardeol a 44: 113-117. Shmida, A. and Wilson, M.V. 1985.Biologica l determinants of species diversity. Journal of Biogeography 12: 1-20. Smith, E.P. and Van Belle, G. 1984.Nonparametri c estimation of species richness. Biometrics 40: 119-129. Steffan-Dewenter, I.Munzenberg , U., Burger, C, Thies, T. and Tscharntke, T. 2002. Scale- dependent effects of landscape structure on three pollinator guilds. Ecology 38: 1421-1432. Stephens, S.E., Koons,D.N. ,Rotella , J.J. and Willey, D.W. 2003. Effects of habitat fragmentation on avian nesting success: arevie w ofth e evidence at multiple spatial scales. Biological Conservation 115: 101-110. Taylor, A.D. 1988. Large-scale spatial structure and population dynamics in arthropod predator-prey systems.Annal e Zoologici Fennici 25:63-74. Taylor, P.D.an d Merrian, G. 1995.Habita t fragmentation andparasitis m of a forest damselfly. Landscape Ecology 11: 181-189. Turner, M.G. 1989.Landscap e ecology: the effect ofpatter n onprocess .Annu . Rev. Ecol. Syst. 29:171-197. Van Dorp,D . and Opdam,P.F.M . 1987. Effects ofpatch size, isolation and regional abundance on forest bird communities. Landscape Ecology (1)1: 59-73. 74 Verboom, B.an d Van Apeldoorn, R. 1990. Effects of habitat fragmentation on the squirrel, Sciurus vulgaris L. Landscape Ecology 4(2/3): 171-176. Vielliard, J.M.E. 2000. Bird community as an indicator of biodiversity: results from quantitative surveys. Brazil.A . Acad. Bras. Ci. 72(3),323-330 . Villard, M.A., Merriam, G. and Maurer, B.A. 1995.Dynamic s in subdivided populations of Neotropical migratory birds in atemperat e fragmented forest. Ecology 76:27-40 . Villard, M.A., Trzcinki, M.K, and Merriam, G. 1999.Fragmentatio n effects of forest birds: Relative influence ofwoodlan d cover and configuration on landscape occupancy. Conservation Biology 13(4):774-783 . Wiegand, T., Revilla, E. and Moloney, K.A. 2005.Effect s ofhabita t loss and fragmentation on population dynamics. Conservation Biology 19(1): 108-121. Wilson, D.S. 1992.Comple x interactions inmetacommunities , with implications for biodiversity and higher levels of selection. Ecology 73: 1984-2000. With, K.A. and Crist, O.C. 1995. Critical thresholds in species' responses to landscape structure. Ecology 76(8):2446-2459 . With, K.A. and King,A.W . 1999. Dispersal success on fractal landscapes: a consequence of lacunarity thresholds. Landscape Ecology 14:73-82 . 75 Annex Order Family Species E PC FG FD Galliformes Cracidae Ortalisvetula NA NA F NA Ciconiiformes Ardeidae Bubulcus ibis NA NA C/I NA Cathartidae Coragypsatratus NA NA C NA Cathartes aura NA NA C NA Falconiformes Accipitridae Buteo magnirostris NA NA C NA Columbiformes Columbidae Patagioenas flavirostris NA NA G/F NA Zenaida asiatica NA NA G/F NA Columbina inca NA NA G NA Columbina talpacoti NA NA G NA Leptotila verreauxi NA NA G/F NA Psittaciformes Psittacidae Pionus senilis NA E F NA Amazona autwnnalis NA NA G/F NA Cuculiformes Cuculidae Piaya cayana NA NA I NA Crotophaga sulcirostris NA NA I NA Strigiformes Strigidae Megascops guatemalae ND NA C NA Glaucidium brasilianum NA NA C NA Ciccaba virgata NA NA C NA Apodiformes Apodidae Streptoprocne zonaris NA NA I NA Chaeluravauxi NA NA I NA Trochilidae Campylopteruscurvipennis NA NA N/I NA Campylopterushemileucurus NA NA N NA Anthracothorax prevostii NA NA N/I NA Hylochoris leucotis NA NA N NA Amazilia Candida NA NA N/I NA Amazilia cyanocephala NA NA N NA Amazilia tzacatl NA NA N/I NA Amazilia yucatanensis MX NA N NA Lampornis amethystinus NA NA N NA Lampornis clemenciae NA NA N NA Eugenes fulgens NA NA N NA Althis heloisa MX NA N NA 76 Order Family Species E PC FG FD Trogoniformes Trogonidae Trogonelegans NA NA 0 FR Coraciiformes Momotidae Momotusmomota NA NA 0 NA Ramphastidae Aulacorhyncus prasinus NA SSP 0 NA Piciformes Picidae Melanerpes formicivorus NA NA O FR Melanerpes aurifrons NA NA I NA Veniliornisfumigatus NA NA I NA Piculus rubiginosus NA NA I NA Dryocopus lineatus NA NA I NA Campephilus guatemalensis NA SSP I NA Passeriformes Dendrocolaptidae Sittasomus griseicapillus NA NA F/I FR Xiphorhynchus flavigaster NA NA I NA Lepidocolaptes affinis NA NA I FR Tyrannidae Camptostomaimberbe NA NA I NA Mitrephanes phaeocercus NA NA I NA Contopus pertinax NA NA I NA Empidonax difficilis NA NA I NA Empidonax occidentalis NA NA I NA Sayornis nigricans NA NA I NA Myiarchus tuberculifer NA NA I/F NA Myiarchus tyrannulus NA NA O NA Pitangus sulphuratus NA NA I/F NA Megarynchus pitangua NA NA 0 NA Myiozetetes similis NA NA I/F NA Myiodynastes luteiventris NA NA I NA Tyrannus melancholicus NA NA 0 NA Tyrannus couchii NA NA O NA Tityra semifasciata NA NA F NA Vireonidae Vireohuttoni NA NA 0 NA Vireoleucophrys NA NA F/I NA Hylophilus ochraceiceps NA SSP I NA Vireolaniusmelitophrys NA NA I NA Cyclarhis gujanensis NA NA I NA 77 Order Family Species PC FG FD Corvidae Cyanocoraxyncas NA NA 0 NA Cyanocorax morio NA NA O NA Hirundinidae Stelgidopteryx serripenis NA NA I NA Hirundo rustica NA NA I NA Troglodytidae Campylorhynchus zonatus NA NA I NA Catherpes mexicanus NA NA I NA Thryothorusmaculipectus NA NA I NA Henicorhina leucosticta NA NA I NA Henicorhina leucophrys NA NA I FR Turdidae Sialia sialis NA NA o NA Myadestes occidentalis NA SSP F NA Myadestes unicolor NA E F NA Catharus mexicanus NA SSP O NA Turdusgrayi NA NA I/F NA Turdusassimilis NA NA I/F NA Mimidae Toxostoma longirostre NA NA F NA Melanotis caerulescens MX NA F NA Parulidae Panda pitiayumi NA NA I NA Geothlypis poliocephala NA NA O NA Euthlypis lachrymosa NA NA I NA Basileuterus culicivorus NA NA I NA Basileuterus rufifrons NA NA I NA Thraupidae Chlorospingus ophthalmicus NA NA O NA Piranga leucoptera NA NA O NA Thraupis abbas NA NA F NA Cyanerpes cyaneus NA NA O NA Emberizidae Volatiniajacarina NA NA G/I NA Sporophila torqueola NA NA G NA Sporophila minuta NA NA G NA Tiaris olivaceus NA NA O NA Atlapetes albinucha NA NA I NA Buarremon brunneinucha NA NA I FR 78 Order Family Species EN PC FG FD Arremonops rufivirgatus NA NA G NA Aimophila rufescens NA NA G NA Spizella passerina NA NA G/I NA Cardinalidae Saltator atriceps NA NA O NA Cyanocompsa parellina NA NA G/I NA Icteridae Dives dives NA NA O NA Quiscalusmexicanus NA NA O NA Molothrus aeneus NA NA o NA Icterus gularis NA NA I/N NA Icterus graduacauda MX NA I/N NA Amblycercus holosericeus NA NA I NA Psarocolius montezuma NA SSP F NA Fringillidae Euphonia hirundinacea NA NA G/F NA Euphonia elegantissima NA NA F NA Carduelis pinus NA NA G NA Carduelispsaltria NA NA G NA Coccothraustesabeillei NA NA G NA Annex 1. Totallist of observed,heard and/or captured birds in taxonomic order inall the landscapes (according toAOU1998). Abbreviations: EN, Endemic; PC, Protection Category;FG, Foraging Guild; FD, Forest Dependent; MX, endemic toMexico; R, resident;M, migratory; E, endangered;SSP, subject tospecial protection; F, frugivorous; I, insectivorous; G, granivorous, N, nectarivorous; O, omnivorous; C,carnivorous; FR, forest dependent (species that breed exclusively in theforest); NA, not applicable. 79 •J- &->. •Trf ' '-l ;,f'' .## S$ i •••^M"'A-- Chapter 4 Influence ofbod y size on coexistence of bird species. Euridice Leyequien, Willem F. deBoer, Antoine M. Cleef Abstract Interspecific competition isthough t tob e an important ecological force structuring ecological communities. Despite almost three decades efforts, there isn o unifying theory that predicts patterns and processes in ecological communities. Body mass is an easily determined characteristic of animals that probably influences competition strength. Our objective ist o examine the effect ofbod y size (mass) on competitive interactions between bird pairs, ranging from hummingbirds species with small body mass tooropendul a species with higher body mass. We hypothesized that the competition strength between competing species will show a negative relationship with body mass ratio,wit h the largest level of competition between species with the lowest body sizeratio .Moreove r species that have a greater overlap inresourc e use tend to exhibit stronger competition than species that overlap less in theirresourc e use.Ou rresult s demonstrate that there is a significant negative relationship between bird body mass ratio and the competition strength i.e.; the largerth ebod y mass ratio,th e lower the competition strength thereby suggesting that high variation inbod y sizes amongst sympatric species may promote coexistence in communities. In addition, we did not find significant influence of foraging strategy type inth e relationship between body mass ratio and competition strength. 82 Introduction Interspecific competition has long been recognised as amajo r ecological force shaping the community patterns (Ricklefs 1975,Gille r 1984,Wiens 1989,Kedd y 2001). However, the mechanisms of such competition have remained elusive. Specifically, one attribute that probably influences competition strength in communities isbod y size of sympatric species. Brown and Maurer (1986) demonstrated that large species dominated smaller species in competition for food. French and Smith (2005) found a significant linear positive correlation between the mass of each species and its dominance rank, which critically influences the resource competition among species.I n addition, Gotelli and Ellison (2002) found that body size in ant communities has important consequences for resource utilisation and species interactions where co-existing species exhibit regular spacing ofbod y sizes. It has been also suggested that species can evade competition by differing inbod y size, as species with dissimilar body masses have dissimilar energetic requirements and capacities in terms of food searching, harvesting, orprocessin g (Bowers and Brown 1982). However, the functional significance ofbod ymas s in community ecologyremain s unclear, and the relationship between coexistence andbod y size differences amongst ecologically similar species is still debated. In addition, it ispresume d that species similar in morphology, physiology, and behaviour will experience more intense competition (Brown and Wilson 1956,Hutchinso n 1959, Grant 1972, Martin and Martin 2001). Particularly, species that exploit the same food resource (i.e.; same foraging strategy) in a similar way will be expected to experience competition, which separates the species ecologically within the region of niche by a guild. (Wiens 1989). 83 Phase portraits Species interactions are the foundation of community ecology. However the identification of species interactions (e.g.; competition) and the quantitative measure of its strength have proven difficult in natural communities. Recently, several studies have concentrated on developing methods that can differentiate among these different "signals", and are able to identify the interaction type and to quantify the interaction strength using time series. Seip (1997)develope d andteste d amethod , called the Key Factor Method, for characterising species interactions from ecological data using time series,plotte d in so-called phase portraits. This method extracts quantitative variables that characterise species interactions. Later an extension of the Key Factor Method, called the Angle Frequency Method was developed and tested in modelling studies (Seip and Pleym 2000, Sandvik etal. 2002 ) and observational studies (Sandvik etal. 2003 , Sandvik etal. 2004) . Both methods are developed from the traditional graphical analysis of species interactions in ecology (Lotka 1925,Volterr a 1926,Rosenzwei g and MacArthur 1963,MacArthu r 1972). The Key Factor Method and the Angle Frequency Method show that species interactions (i.e.;prey-predation , competition, mutualism and facilitator-gainer) could all be distinguished by their biomass trajectories, where different types of ecological interactions will show unique patterns inphas e portraits. This simple approach avoids past difficulties where studies relied on simplifying assumptions which are often violated in ecological synoptic data, orth e need ofextensiv e information aboutth e studied community (Sandvik etal. 2004).I n an ideal competition interaction, an increase of species a will cause a proportional decrease of species b.Th e corresponding phase portrait (i.e.;wher e the biomass of both species are plotted on thex and>> axis) will show a sequences of trajectories along a line through the centre point at a 135° angle to thex axis, clearly distinctive from other interaction types (i.e.;mutualism , prey-predator and facilitator- gainer) (Gilpin etal. 1982, Seip 1997). 84 Moreover, even though a diverse oftheorie s predict the conditions inwhic h species interactions ought to be strong (Sarnelle 1994,Thompso n 1999), itha sprove n to be difficult to determine the actual strength in a quantitative approach. Despite these difficulties, the aforementioned methods showed to be successful in differentiating competition interactions and the strength ofth e interactions in natural communities. Our objective ist o examine the effect ofbod y size on competitive interactions between species pairs of ecologically similar birds. We hypothesized that the competition strength between competing species will show anegativ e relationship with body size ratio,wit h the largest level of competition between species with the lowest difference in body size. Moreover we will test whether species that share food resources (i.e.;belongin g to the same foraging guild) will exhibit stronger competition than species that differ in their resource use. Materials and Methods The bird species data setwa s a compilation of presence-absence and abundance data obtained from point count observations from the Cuetzalan Region (~ 54,200 ha) in the north-eastern mountain range of Puebla, Mexico, from November 2002 toNovembe r 2003 (Leyequien etal. submitted). We used standard point count techniques for bird detection (Bibby etal. 1992),wit h±1 0 point counts, replicates per sitepe r month. A modified double-observer approach was used for visual and sound avian detection (Parker 1991). Bird observations were recorded over 360° with a fixed-radius of2 5 m, and a counting period (single count) often minutes (Dawson etal. 1997;Vielliar d 2000). A distance of 100m between point counts was used to avoid dependency between points' data. From this species data set we calculated the average number of birds for each species for each of the 12months . Selection ofspecies and construction of the competition prototype. From the bird species data setw e selected 20 competitive species pairs and six non competitive (see Table 1 for species list).Th e competitive species pairs were selected based 85 on body mass (ranging from 6.2 g to 324 g) and foraging strategy, i.e.; (a) diet specialists, (b) opportunistic feeders which depends on aprimar y food source other than the secondary food source, and (c) generalists taking a wider range of diet. For the non-competitive pairs, species were selected based on body mass but with unrelated foraging strategies. To compare and estimate the similarity of these observed pairs to a theoretically ideal competition interaction, we constructed aprototyp e case where the assumptions underlying the expected patterns for competition were fulfilled, i.e.; competing species replace each other consequently; an increase inth e biomass of one species causes a proportional decrease in the biomass of the species (Seip, 1997).I n addition, we constructed purely random interactions to identify stochasticity in the pairwise interactions. The random interaction was constructed using random numbers created by a random number generator with 10.000 permutations. Table 1.Bird species selectedfor pair-wise comparison. Status (S)code: R,resident, ;M, migratory. Diet code: I, insectivore;F, frugivore; N, nectarivore; G, granivore; O, omnivore. Common Name Scientific name Order Family S Diet Red-billed Pigeon Columba flavirostris Columbiformes Columbidae R G/F White-tipped Dove Leptotila verreauxi Columbiformes Columbidae R G/F White-crowned Parrot Pionus senilis Psittaciformes Psittacidae R F Squirrel Cuckoo Piaya cayana Cuculiformes Cuculidae R I Wedge-tailed Sabrewing Campylopteruscurvipennis Apodiformes Trochilidae R N/I White-bellied Emerald Amazilia Candida Apodiformes Trochilidae R N Blue-crowned Motmot Momotusmomota Coraciiformes Momotidae R 0 Golden-fronted Woodpecker Melanerpes aurifrons Piciformes Picidae R I Lineated Woodpecker Dryocopus lineatus Piciformes Picidae R I Dusky-capped Flycatcher Myiarchus tuberculifer Passeriformes Tyrannidae R I/F Great Kiskadee Pitangus sulphuratus Passeriformes Tyrannidae R 1/F Social Flycatcher Myiozetetes similis Passeriformes Tyrannidae R I/F Sulphur-bellied Flycatcher Myiodynastes luteiventris Passeriformes Tyrannidae R I Masked Tityra Tityra semifasciata Passeriformes Tyrannidae R F Green Jay Cyanocorax yncas Passeriformes Corvidae R O 86 Common Name Scientific name Order Family S Diet Brown Jay Cyanocorax morio Passeriformes Corvidae R O Band-backed Wren Campylorhynchuszonatus Passeriformes Troglodytidae R I Spot-breasted Wren Thryothorus maculipectus Passeriformes Troglodytidae R I White-breasted Wood-Wren Henicorhina leucosticta Passeriformes Troglodytidae R I Blue-gray Gnatcatcher Polioptila caerulea Passeriformes Sylviidae M I Clay-colored Robin Turdusgrayi Passeriformes Turdidae R I/F White-throated Robin Turdusassimilis Passeriformes Turdidae R I/F Black-throated Green Dendroica virens Passeriformes Parulidae M I Wilson's Warbler Wilsoniapusilla Passeriformes Parulidae M I Golden-crowned Warbler Basileuterus culicivorus Passeriformes Parulidae R I Rufous-capped Warbler Basileuterus rufifrons Passeriformes Parulidae R I Yellow-throated Euphonia Euphonia hirundinacea Passeriformes Thraupidae R G/F Yellow-winged Tanager Thraupisabbas Passeriformes Thraupidae R F Green-backed Sparrow Arremonops ruflvirgatus Passeriformes Emberizidae R G Dark-backed Goldfinch Carduelis psaltria Passeriformes Fringillidae R G Melodious Blackbird Dives dives Passeriformes Icteridae R O Great-tailed Grackle Quiscalus mexicanus Passeriformes Icteridae R O Altamira Oriole Icterus gularis Passeriformes Icteridae R I/N Black-headed Oriole Icterus graduacauda Passeriformes Icteridae R I/N Montezuma Oropendola Psarocolius montezuma Passeriformes Icteridae R F Form factors We used the key factor method to characterise the species pairwise interactions (Seip, 1997).First , we constructed phase portraits for each species pair, including the random interaction and the ideal competitive pairs,wher e the average number of individuals for species x and y was plotted using the 12synopti c samples. Thephas e portrait for the prototype competition interaction wasplotted ; the sequences oftrajectorie s are depicted as a line through the centrepoin t at a 135° angle to the x axis (Gilpin etal. 1982,Sei p 1997, Seip and Pleym 2000).Thereafter , we extracted 13for m factors from each of the phase 87 portraits of pairwise interactions. The form factors express attributes that characterise each phase portrait for competition (see Seip, 1997). To calculate dissimilarity scores (distances) for each ofth e pairwise interactions with the prototype competition and random pairs we used clustering techniques based on principal component analysis (PCA) (Jongman et al. 1987). We constructed a matrix (27x13) that contained the independent form-factor variables, with 25 species pairs, one random interaction and one prototype competition interaction (rows), and 13for m factor parameters (columns). The scaling was done by inter-species correlations and the species scores were divided by standard deviations; data was not transformed, and centred by species. The resulting scatter plot shows points representing the interacting pairs in a coordinate system formed by the principal axes defined by the PCA. We used the scores ofth e first and second axis to calculate distance scores (D) from the interacting pairs to the prototype competition interaction. We used the Pythagorean Theorem to calculate the distance between each ofth e interacting pair (xy) and the reference competition point from their coordinates on the first and second principal component axis. These distance scores (D) were used as the dependent variable for hypothesis testing. To test ifther e were significant differences between the means ofth e random and potentially competitive pairs,w e used a One-WayANOV A procedure, with the dependent variable D, and the random/non-random pair interactions as independent factor. Next, we calculated the ratio in Inbod y mass (InBM) for each ofth e interacting pairs body masses of thebir d species were taken from Dunning (1993). Subsequently, we calculated the association between the InBM ratio and the strength of competition (D) using a Pearson product-moment correlation coefficient (SPSS 12). Subsequently, we used aunivariat e general linear model (GLM with Type III sums of squares) for regression analysis, where the InBMratio s were used as the covariate and the distance scores D asth e dependent variable. We included two fixed factors to categorize the species pairs. The first fixed factor contained three groups, classifying the different foraging guilds into specialist-generalist classes. The second fixed factor was used to distinguish the classes for random/non-random pair interactions. 88 Results Competition dissimilarity scores The PCA scores for the assumed competitive and random interactions were dissimilar, forming distinct clusters, indicating signals of competition or stochasticity for the different species pairs (Figure 1).Th e first principal component axis explained 50%o fth e variability inth e data, and the second axis explained 43%, totalling 93 % for the two axes. Figure 1. Thescore plot shows thefirst principal component inthe x-axis, which explains 50% of the variance, and thesecond principal component inthey-axis, which explains 43% of the variance. It shows theposition of theprototype interactionfor competition (COM), thestochastic interaction (STO), the interactingpairs (Cj. . .C,) and the randompairs (Rj. ••Rn>- The analysis ofvarianc e showed that the D scores of the random and competitive groups = were significantly different (Fi>26 7.226, P = 0.015) (Fig. 2),wit h a higher D score (i.e.;a larger distance from the competition reference point) for random pairs compared to potentially competing pairs. 89 Stochasticity Figure 2. Significant differences inthe meanD scores of the two random and competitive groups. 1 = Random, 0 = Competitive. Effects of bodymass onspecies competition We found a significant negative correlation between the body mass ratio and the strength of competition (D) (r= 0.536139;P = 0.02,R 2= 0.446) (Figure 3),not e that they axis ranges from strong to weak competition starting from the origin (0).I n addition, the foraging guild included as a fixed factor in the model did not show a significant effect of foraging guild on competition strength (D). 90 2.5 Figure 3.Positive correlation between the body mass ratio, i.e. \lnBMspeciel- lnBMspecies2\, and thestrength of competition (D).Here they axis ranges strong to weak competition startingfrom theorigin (0). Discussion We have demonstrated a significant negative relationship between bird body mass and the competition strength; the largerth e body mass difference between species pairs the lower the competition strength. This isn o direct proof that species coexistence in communities is structured by difference in body masses,bu t evidence of size-related resource division sustains the conclusion that different size classes inth e community may promote coexistence (Hutchinson 1959,Bower s andBrow n 1982,Gran t 1986,Ernes t 2005). The encountered relationship may alsoresul t from restrictions imposed by size-dependent metabolic rates,bioti c interactions, and energy availability to different size classes (Knouft 2002). Previous results have suggested that body size-mediated competition may be important for structuring energy use (Ernst 2005), and that energy isunequall y available 91 across body sizes. Therefore body sizes with the greatest access to resources would be favoured, resulting in species aggregations around specific masses (Holling 1992). Moreover, Brown and Bowers (1984) found that species of similar size (body mass) coexist less frequently in local communities suggesting that their co-occurrence ispreclude d by interspecific competition. Our results indicate that similarity in body size (i.e.;a low body mass difference) increases competition, and thereby would promote character displacement. However, no data are available about the resource availability, sotha t itremain s unclear whether the coexistence of species with relative similar body sizes issupporte d by unequal access toresources . In addition, morphological differences (i.e.;bea k size and structure) in similar sized bird species that occupy the samehabita t might promote coexistence by feeding niche differentiation (Grant 1966, 1968, 1999,Baldwi n 1953,Free d etal 1987,Conan t 1988). In addition, in accordance with our results, itha s been suggested that body mass isth e main factor influencing the dominance patterns of competitive species rather than diet specialisation (French and Smith 2005). In our results we did not find a significant influence of guild type in the relationship between body mass and competition strength. Alternative explanations are based on the theory of speciation and radiation where genetically variable traits in morphological differences (e.g.;bea k size) mayresul t in greater specialisation in obtaining someparticula r food, and therefore avoiding competition (Grant 2001). Besides, itha s been suggested that allometric scaling partly determines coexistence through competition, thereby structuring species assemblages, especially between closely related species of similar size in the same guild (Brown and Bowers 1984). Moreover, our results indicate that the body mass-competition strength relationship may be a valuable robust and powerful tool to identify patterns ofcommunit y structure. However, further exploration ofth e body mass-competition relationship, considering different temporal and spatial scales, as well as different taxa should produce further insights into the forces shaping community structure and species coexistence. Moreover, correlative- 92 experiments that measure the competition intensity should beperforme d to corroborate existing gradient in competition intensity (i.e.;fro m diffuse competition to monopolistic competition) in the natural community. Acknowledgments We would like to acknowledge Knut L. Seip for his advice and comments, and Dr. Ir. Andre Schaffers for statistical consultation. Namgail Tsewang provided comments that improved this manuscript. This research was supported by the CONACYT grant 149557 to E. Leyequien 93 References Baldwin, P.H. 1953. Annual cycle,environmen t and evolution inth e Hawaiian honeycreepers (Aves:Drepaniidae) .Univ .calif .Publ .Zool . 52: 285-398. Bibby, C.J., Burgess, N.D. and Hill, D.A. 1992. Bird census techniques. Academic Press, California. Bowers,M.A . and Brown, J.H. 1982.Bod y size and coexistence in desert rodents: chance or community structure? Ecology 63(2):391-400 . Brown, J. H., and M. A. Bowers. 1984. Patterns and processes in three guilds of terrestrial vertebrates. Pages 282-296. In: Strong, Jr., D. R. , Simberloff, D., Abele, L. G. and Thistle, A. B. (eds.) Ecological communities: conceptual issues and the evidence. Princeton University Press, Princeton, N.J. Brown, J.H. and Maurer, B.A. 1986.Bod y size, ecological dominance and Cope's rule. Nature, 324: 248-250. Brown, W.L. and Wilson, E.O. 1956.Characte r displacement. Systematic Zoology 5:49 - 64. Conant, S. 1988.Geographi c variation inth e Laysan Finch (Telespiza cantans).Evol .Ecol .2 : 270-282. Dawson, D.K., Smith, D.R. and Robbins, C.S. 1997. Point count length and detection of forest Neotropical migrant birds. In: Ralph, C.J., Sauer, J.R. and Droege, S. Monitoring Bird populations by point counts. General technical Report PSW- GTR-149. Albany, California. Dunning, J. B., Jr. 1993. CRC handbook of avian body masses. CRC Press, Boca Raton, Florida, USA. Ernest, S.K.M.2005 .Bod y size, energy use, and community structure of small mammals. Ecology 86(6): 1407-1413. Freed, L., Conant, S. and Fleischer, R.C. 1987. Evolutionary ecology and radiation of Hawaiian passerine birds.Trend secol .Evol .2 : 196-203. French, A.R. and Smith, T.B. 2005.Importanc e of body size in determining dominance hierarchies among diverse tropical frugivores. Biotropica 37(1): 96-101. 94 Giller, P.S. 1984.Communit y structure and the niche. Chapman and Hall, London. Gilpin, M.E., Case, T.J. and Bender, E.A. 1982.Counterintuitiv e oscillations in systems of competition and mutualism. American Naturalist 119(4): 584-588. Gotelli,N.J . and Ellison, A.M. 2002.Assembl y rules for New England ant assemblages. Oikos 99:591-599 . Grant, P.R. 1966. Ecological compatibility of bird species in islands. American Naturalist 100:451-462. Grant, P.R. 1968. Bill size, body size, and the ecological adaptations of bird species to competitive situations on islands. Syst.Zool . 17:319-333 . Grant, P.R. 1972.Convergen t and divergent character displacement. Biol.J . ofth e Linnean Soc.4:39-68 . Grant, P.R. 1986.Interspecifi c competition, island biogeography and null hypotheses. Evolution 34: 332-341. Grant, P.R. 1999. Ecology and evolution of Darwin's finches, 2nd ed., Princenton University Press. Grant. P.R. 2001. Reconstructing the evolution of birds on islands: 100 years of research. Oikos 92:385-403 . Holling, C.S. 1992.Cross-scal e morphology, geometry, and dynamics of ecosystems. Ecological Monographs 62:447-502 . Hutchinson, G.E. 1959.Homeag e to santa rosalia; or, why arether e soman y kinds of animals? American Naturalist 93: 145-159. Jongman, R.H.G., TerBraak , C.J.F. and Van Tongeren, O.F.R. 1987. Data analysis in community and landscape ecology. Cambridge University Press, Cambridge. Keddy, P.A. 2001.Competition . KluwerAcademi c Publishers, Dordrecht. Knouft, J.H. 2002.Regiona l analysis of body size and population density in stream fish assemblages: testingprediction s ofth e energetic equivalence rule. Can. J. Fish. Aquat. Sci. 59: 1350-1360. Leyequien, E., De Boer, W.F.an d Skidmore,A.K . (submitted) Linking species- environment relationships and multiple spatial scales in community ecology. 95 Lotka, A. J. 1925. Elements of physical biology. Williams and Wilkins, Baltimore, Maryland, USA. MacArthur, R.H. 1972. Geographical ecology. Harper and Row,Ne w york,Ne w York, USA. Martin, P.R. and Martin, T.E. 2001.Ecologica l and fitness consequences of species coexistence: a removal experiment with wood warblers. Ecology 82(1): 189-206. Owen-Smith, R.N. 1992. Megaherbivores. The influence of very large body size on ecology. Cambridge studies in ecology. Cambridge University Press, Cambridge. Parker, T.A. 1991.O n the use of tape recorders in avifaunal surveys. The AUK 108(2):443- 444. Ricklefs, R. 1975.Competitio n and the structure ofbir d communities. Evolution 29: 581- 584. Rosenzweig, M.L. and MacArthur, R.H. 1963. Graphical Representation and Stability Conditions of Predator-Prey Interactions. American Naturalist 97:209-223 . Sandvik, G., Seip,K.L . and Pleym, H. 2002.A n anatomy of interactions among species in a seasonal world. Oikos 99: 260-271. Sandvik, G., Seip, K.L. and Pleym, H. 2003. Extracting signals of predation and competition from paired plankton time series. Arch. Hydrobiol. 157(4): 455-471. Sandvik, G., Jessup, CM., Seip, L.S. and Bohannan, B.J.M. 2004. Using the angle frequency method to detect signals of competition and predation in experimental time series. Ecology Letters 7:640-652 . Sarnelle, O. 1994. Inferring process from pattern: trophic level abundances and imbedded interactions. Ecology 75: 1835-1841. Seip, K.L. 1997.Definin g and measuring species interactions in aquatic ecosystems. Can. J. Fish. Aquat. Sci. 54: 1513-1519. Seip, K.L. and Pleym, H. 2000. Competition and predation in a seasonal world. Verh. Internat. Verein. Limnol. 27: 1-5. Thompson, J.N. 1999.Th e evolution of speciesinteractions . Science 284:2116-2118 . 96 Vielliard, J.M.E. 2000. Bird community as an indicator of biodiversity: results from quantitative surveys. Brazil. A. Acad. Bras. Ci. 72(3): 323-330. Volterra, V. 1926. Fluctuations in the abundance of a species considered mathematically. Nature 118:558-560. Wiens, J.A . 1989b. The ecology of bird communities. Volume 2. Processes and variations. Cambridge University Press. 97 Chapter 5 Linking species-environment relationships and multiple spatial scales in community ecology Euridice Leyequien, Willem F. deBoer, Andrew K. Skidmore Abstract Biological communities are organised at multiple functional spatial scales and interactions between these scales determine both local and regional patterns of species richness. Despite the recognition that species-environment relationships are scale-dependent many ecologists have neglected the influence of scale on species richness patterns andprocesses . We analysed the influence of spatially explicit bio-physical variables on abir d community in a Mexican region denominated as an important area for bird conservation. Using a multiple scale approach with plot, patch and landscape level variables using abundance and presence-absence data, we demonstrated that landscape variables explained most ofth e variation inbir d species inbot h abundance and presence-absence analyses in all explanatory sets.Interestingly , results demonstrated that variation incommunit y structure was described best at family-level than at genera- or species-level. In addition, shade coffee plantations could provide habitat for neo-tropical migrants and forest-dependent birds (e.g.; endemic,protecte d species). Selecting the appropriate scale(s) of management in conservation strategies could have important implications for conservation ofbir d communities in Cuetzalan region. 100 Introduction There is increasing recognition that community structure, species richness and abundances may depend on multiple spatial and temporal scales. Because ecological patterns, processes, and species responses are scale-dependent, issues of scale are central to ecology (Meentemeyer and Box 1987; Wiens 1989; Levin 1992, 1993;Dal e et al. 1994; Bissonette 1997). Besides, biological communities are organised at multiple functional spatial scales and interactions between these scales determine both local and regional patterns of species richness (Loreau 2000; Mouquet etal. 2003; Kneitel and Chase 2004; Borcard et al.2004) . Understanding the species-environment interactions and how this affects the richness and abundance of biological communities will require integrating information from scaling effects in ecology. Despite the recognition that diversity patterns are scale-dependent in species-environment relationships, many ecologists have neglected the influence of scale on these patterns and processes. Studies addressing relationships between community patterns and single causes at single hierarchical levels can be misleading (Wiens 1989; Borcard et al. 1992; Cushman and McGarigal 2002). Traditional community studies usually compare the marginal effects of factors at single scale. However, understanding relationships between community structure and environmental factors isgreatl y benefited by partitioning species-environment relationships to quantify the magnitude of the independent and confounded effects of each component (Cushman and McGarigal 2002). Recently, approaches to explain species diversity patterns using environmental variables at multiple scales have been attempted, where the independent and confounded effects are partitioned (Borcard et al. 1992; Liu and Brakenhielm 1995; Cushman and McGarigal 2002; Borcard et al. 2004; Cushman and McGarigal 2004). Despite these efforts, there is a lack of multi-scale and hierarchical approaches to understand the patterns in ecological communities' organisation, and the processes and causal mechanisms that act as driving forces. 101 In this paper we investigate the influence of spatially explicit bio-physical variables on a bird community at different taxonomic levels, incorporating different class groups (i.e.; migrants, endemics and species with special protection status), using a multiple scale approach applying three scales:plot ,patch and landscape. We presented our approach using the data from a bird community study in Cuetzalan region, Mexico. We studied shade coffee plantations, a habitat with large ecological benefits, widely documented by many researchers (Greenberg et al. 1995; Moguel and Toledo 1998; Gobbi 2000; Liang et al. 2001; Soto-Pinto et al. 2000). We analysed both species abundance and presence-absence data because the different analysis may produce different results regarding the relative influence of species-environmental relationships among various spatial scales, as recoding from abundance to presence-absence data removes information regarding the ecological relationships between species abundance patterns and environmental gradients (Cushman and McGarigal 2004). Moreover, it has been suggested that descriptors of landscape structure and composition are stronger predictors of bird species assemblages and community composition (O'Neill et al. 1997; Miller et al. 2004). We therefore hypothesised that landscape-scale explanatory variables would be stronger predictors of bird species assemblages than patch-scale orplot-scal e variables. The variation in community structure, furthermore, is often sensitive to the number of variables (i.e.; taxa) used to describe each observation (i.e.; community) (Bouwman and Bailey1997) . Most studies deal with species-level data to predict diversity patterns; however family-level data have the advantage that could facilitate interpretation of sources and sinks of biodiversity along spatial scales. Moreover, the distribution of whole families is often less complex, facilitating interpretation of the mechanisms leading to the observed patterns (Jacobsen 2004). It is therefore relevant to identify the taxonomic level at which the variation in community structure is best explained across different spatial scales. Thus, our second hypothesis is that the variation in community structure will be best described at family-level than genera- or species-level independent ofth e spatial scale of the analysis. 102 Methods The species data set consisted of 170 bird species recorded in nine landscape frames of 10x 10 km in 120 visits during the migratory and breeding seasons from November 2002 to November 2003 in the North Eastern Mountain range of Puebla, within the Cuetzalan region, Mexico. We used standard point count techniques for bird detection (Bibby et al. 1992). A number of ±10 point counts per site were performed. A modified double-observer approach was used for visual and sound avian detection (Parker 1991). Bird observations were recorded using over 360° with a fixed-radius of 25 m, and counting period (single count) often minutes (Dawson et al. 1997; Vielliard 2000). A distance of 100 m between point counts was used to avoid dependency between points' data. The bio-physical data set consisted of three sub-sets of variables at three different scales: plot (0.025 hectares, n=36), patch (>1 to <3 km, n=12) and landscape (radius of 1, 3,5 , and 5-10 km, n=28). The plot sub-set included floristic, stand structure and vegetation cover variables (Table 1). The patch sub-set included patch configuration, human disturbance, water resource availability, Normalised Difference Vegetation Index (NDVI) average (as a proxy for vegetation vigour), average slope, and altitude variables. Subsequently, we derived a landscape sub-set that consisted of landscape composition metrics measured for each landscape frame at three different radiuses. We extracted variables at plot-scale insitu from 0.025 ha plots located in shade coffee plantations in each of the nine sampled landscape frames using standard quadrants (Kent and Coker 1992). For patch and landscape variables, we computed a multi-spectral image classification using an ETM satellite image (year 2003; 30 meters of spatial resolution). We applied a maximum likelihood classifier method, including prior probabilities to generate a land cover map, which contained seven classes (i.e.;agricultura l land, grazing land, coffee plantations, forest, secondary vegetation, urban areas and water bodies). For patch-scale variables (i.e.;patch defined as the polygon containing bird point counts) we derived patch-configuration properties. Human disturbance variables were calculated using distance operations. As for distance to human settlements we used an increasing radius from 500, 1000, 1500 and 2000 meters, computed 103 from the central point of each patch. Additionally, we calculated distance to primary or secondary roads, and also distance to water resources using the shortest linear path in any direction. Vegetation vigour variables were derived from NDVI values computed from a two-band combination index (Lillesand and Kiefer 1979). In addition, for altitude and slope variables, average values of each patch were calculated. In all patch variables calculations we used ILWIS 3.2 (Integrated Land and Water Information System, 1997). Table 1 Explanatory variablescontained byplot, patch and landscape sets, bothfor abundance andpresence-absence analyses. Explanatory set Explanatory variable class Description Plot Floristic Total number oftree species in aplo t of 0.1h a per site within coffee plantations Total number of herb species in a plot of 0.1 ha per site within coffee plantations Stand structure Standard deviation of overall tree heights Average height of coffee shrubs (m) in a plot of 0.1 ha per site Calculated density of tree species per hectare within coffee plantations Average number of coffee individuals in a plot of 0.1 ha per site within coffee plantations Vegetation cover Total amount of crown spread of canopy in a plot of 0.1 ha per site within coffee plantations Standard deviation of crown spread of canopy in a plot of0. 1 hape r site within coffee plantations Average crown spread of canopy in a plot of 0.1 ha per site within coffee plantations Patch Configuration Area (ha) of coffee plantation patch in which plot was located Perimeter of coffee plantation patch in which plot was located Disturbance factors Distance from patch to human settlements in an increasing radius of 500, 1000, 1500 and 2000 m (total of 4 variables per patch) Distance from patch to the nearest primary or secondary road (m)pe r site Water availability Distance from patch toth e nearest water body (m)pe r site Vegetation vigour Average of the Normalised Difference Vegetation Index for patch Standard deviation of the Normalised Difference Vegetation Index for patch Physical (abiotic) Average elevation for bird point counts contained ina patch Average percentage of slope for bird point counts contained in a patch Landscape Landscape composition Proportion of cp, f, sv, u, ag, g, w land covers (ha) with an increasing radius from 1, 3, 5 and >5 kilometres (total of 28 variables). Abbreviations: cp, Coffee plantations; f, Forest; sv, Secondary vegetation; u, urban settlements; ag, Agricultural area; g, grazing land; w, water bodies. 104 We conducted a series of multiple linear regressions with forward selection of independent variables for each species with the zero-truncated abundance data (a = 0.05). Zero counts were only used in the logistic regression. Abundance data were checked for normality, and untransformed abundance data used in the analysis. First, we partitioned the total variance (ft2) of the species data, for each of the 114 species, explained for by the explanatory variables of seven different sets representing the individual and confounded influence of variables at plot, patch and landscape scales (Figure 1). This latter was calculated for the total number of species models and also for only those withP < 0.05. Figure 1. Explanatory sets representing the individual and confounded influence of variables at plot, patch and landscape scales. Abbreviations: Ii, Individual influence; Ci, confounded influence (adaptedfrom Cushman andMcGarigal 2004). In addition, we calculated the total percentage ofbir d species with models withP < 0.05 for all individual explanatory sets and for those representing the influence of the largest explanatory sets. We also calculated the percentage of species for which each of the independent variables had an influence (P <0.05) , so that we could filter out the relative importance of specific independent variables within an explanatory set. However, the 105 number of models with P <0.0 5 increases when variables are added to the model due to spurious correlations. We therefore compared the observed number of species models against the expected number, calculated from a set of random variables (created by a random number generator) regressed against species, with 10,000 permutations. Finally, we grouped the species data to genera and families, and estimated the percentage of models at family and genera level with P <0.05 , for all explanatory sets. This approach enabled the comparison of the effect of taxonomic resolution on the analysis. A similar comparison of the multiple regression performance was executed to compare different species groups (resident-migrants, endemics and species with protection status). For presence-absence analyses we used a series of binary logistic regression with LR forward selection of independent variables, and consequently used the same modified variance partitioning approach, as used for the abundance analyses to quantify the variation (Nagelkerke ^?2) in bird species presence explained by the single and confounded explanatory sets. Finally, we also checked for spurious correlations comparing the observed number of species models against the expected number following the same method as for abundance analyses. Results The explained variance (R2) for the regression models for species abundance analysis varied between 0.43-0.60, whereas when only considering statistically significant models (P < 0.05) the range of explained variance decreased to 0.39-0.54 (Table 2). Results indicated that for statistically significant models, the total variance explained by the Landscape explanatory set individually (0.52) shows a marginal improvement compared to the Patch set (0.51), whereas compared to the Plot set (0.39) is much higher. The Landscape-Plot explanatory set increased the fraction of explained variance to 0.65. The largest Landscape-Patch-Plot set explained 54 % of the total variation, whereas Landscape- Patch set resulted in a total R2 =0.50 . The marginal effect of the Landscape component in all explanatory sets was significantly higher than for the plot or patch subsets. The conditional effect of all components in each different explanatory sets had lower explanatory power than any individual component with the exception of the Patch-Plot set 106 where the conditional effect was higher than the marginal effect of Patch or Plot. The explanatory sets for all models showed a general higher explained variance compared to only statistically significant models,wit h the exception of LS-Plot, LS and Patch sets. The total percentage of species that presented regression models with P < 0.05 was between 54-65 % for all explanatory sets. The highest percentage of bird species that presented regression models with P < 0.05 was accounted for by the confounded influence of Landscape-Patch (65 %), whilst within this explanatory set, the highest percentage of species was obtained by the Landscape variables. However, in comparison to the expected values, the Landscape set presented less regression models with P < 0.05 (62%) than expected on the basis of spurious correlations alone (76%). The cause of this low explanatory power after accounting for spurious correlations could be the result of a large number of landscape variables in the analysis. Table2 Results of theanalyses of birdspecies abundance using multiple linear regressions, utilisingforward selection for different explanatory sub-sets landscape, patch and plot. Mean R for all species and R2for only thespecies thatpresented regression models with P < 0.05 are listed. Additionally, thepercentage of species thatpresented regression models with P < 0.05 is given (N = 114), together with the expected percentage from spurious correlations. Also thepercentage ofspecies thatpresented regression models with P < 0.05 for each of the explanatory sets is given. * P < 0.005; Expected. Abbreviations: Landscape (LS), Patch (PT)and Plot (PL). Total variance Variance explained (%)* All species Model % species*1 Individual Confounded Set R2 R2 LS PT PL LS/PT LS/PL PT/PL LS/PT/PL LS-PT-PL 0.60 0.54 64 (92) 25 11 4 5 5 3 1 LS-PT 0.59 0.50 65 (87) 30 15 - 5 - - LS-PL 0.50 0.65 62 (85) 47 - 12 - 7 - PT-PL 0.55 0.14 63 (66) - 0.9 0 - 14 - LS 0.46 0.52 62 (76) 52 - - - - - PT 0.48 0.51 55 (46) - 51 - - - - PL 0.43 0.39 54(37) - - 39 - - - 107 A more appropriate way to study the impact of individual variables is to calculate the average number of regression models (P <0.05 ) per variable for the different sets (Figure 2). Especially, in the larger spatial scale Landscape, explanatory variables produced more significant results than Patch or Plot variables. The percentage of regression models (P < 0.05) were significantly higher when increasing the spatial extent of the variables in the Landscape set (Spearman-r = 0.587, P < 0.001, N— 28). The most significant variables in Landscape-Patch set were landscape composition (46 % of the species with significant models; Table 3) and human disturbances (8 %). Landscape configuration, vegetation cover and human disturbance variable groups accounted for the majority of the significant regression models inth e other explanatory sets. D Linear U Logistic Figure 2. The average percentage (±s.d.) of models with P < 0.05,per variablesfor the spatial scales at which the variables were measured: Plot, Patch and Landscape sets (<1 km, <3 km, <5 km, and 5-10 km)for both the multiple linear regression and binary the logistical regression. 108 Table 3 Results of the analyses of bird species presence-absence using binary logistic regressions, utilisingforward selectionfor different explanatory sub-sets landscape, patch andplot. Mean Nagelkerke R2for all species, and mean Nagelkerke R for only the species thatpresented models with P <0.05 are listed. Additionally thepercentage of species that presented models with P <0.05 isgiven (N = 114), together with the expected percentage from spurious correlations, as well as thepercentage of species thatpresented models with P < 0.05 for each of the explanatory sets* P < 0.005; (1) Expected. Abbreviations: Landscape (LS),Patch (PT)and Plot (PL). Total variance Variance explained (%)* explained All models Models * Marginal Conditional Set % species* 1 R2 R2 LS PT PL LS/PT LS/PL PT/PL LS/PT/PL LS-PT- 71 (92) 0.68 0.69 26 16 6 8 8 4 0.9 PL LS-PT 70(87) 0.67 0.68 35 19 - 14 LS-PL 67(85) 0.61 0.62 40-11 - 11 PT-PL 66(66) 0.49 0.48 - 21 11 - - 16 LS 63(76) 0.58 0.58 58 PT 63(46) 0.42 0.42 42 PL 61 (37) 0.30 0.34 34 For presence-absence analysis the explained variance (Nagelkerke R2) varied between 0.0.30-0.68, whereas when only considering statistically significant models (P <0.05 ) the range of explained variance increased to 0.34-0.69. For only statistically significant models, the total variance explained by the Landscape (LS) explanatory set (0.58) was much higher than for Patch or Plot sets. The largest Landscape-Patch-Plot set had a Nagelkerke R2 of 0.69, showing a marginal improvement compared to the Landscape-Patch set (Nagelkerke R2 = 0.68). The Landscape-Plot and Patch-plot explanatory sets decreased the fraction of explained variance to 0.62 and 0.48 respectively. Similar to the species abundance analysis, the marginal effect of the Landscape sub-component in all explanatory sets was significantly higher than for the Patch and Plot subsets. The conditional effect of all components in each different explanatory sets had a lower explanatory power than any individual component with the exception of the Patch-Plot set where the conditional effect 109 was higher than the marginal effect of Patch and Plot. For all the models, the explained variance for the largest LS-Patch-Plot showed a marginal improvement compared to LS- Patch and LS-Plot, whereas compared to Patch-Plot is much higher. The LS explanatory set explained a higher variance than Patch or Plot. In the presence-absence analysis, the percentage of statistically significant models (P < 0.05) varied between 61-71 % for the different explanatory sets. In all explanatory sets, Landscape explanatory set accounted for the highest percentage of statistically significant models (P <0.05 ) (Table 4). However, also here the patterns are concealed by the different number of variables used in the models. The Landscape setpresente d a lower percentage of models with P < 0.05 than expected, whereas the Patch and Plot sets produced a higher number of models with P < 0.05 than expected from spurious correlations. The most significant variables were also landscape configuration presenting the highest percentages of models with P < 0.05 of all explanatory sets, followed by human disturbance, stand structure and vegetation cover variable groups (Table 5). Table 4 Percentage of species (N = 114) using abundance analysis (P <0.05) for each of the different variablesper explanatory set. Abbreviations: Landscape (LS),Patch (PT) and Plot (PL), Configuration (C), Abiotic (A), Human Disturbance (HD), Water availability (WA), Vegetation vigour (W), Floristic (F), Stand Structure (SS) and Vegetation Cover (VC). Partial contribution of independent variable group (%) Set LS variables PT variables PL variables LSC PTC A HD WA VV F SS VC LS-PT-PL 48 2 5 11 0 5 4 8 15 LS-PT 46 0.9 7 8 0 5 - - LS-PL 46 - - - - - 3 6 12 PT-PL - 4 9 10 5 13 8 10 26 #variables per class 28 2 2 4 1 2 2 4 3 110 Table 5 Percentage of species (N = 114) usingpresence-absence analysis (P < 0.05) for each of the different variables per explanatory set. Abbreviations: Landscape (LS), Patch (PT) and Plot (PL), Configuration (C), Abiotic (A), Human Disturbance (HD), Water availability (WA), Vegetation vigour (W), Floristic (F), Stand Structure (SS) and VegetationCover (VC). Partial contribution of explanatory variables group (%) Set LS variables PT variables PL variables LSC PTC A HD WA W F SS VC LS-PT-PL 49 3 4 16 0 8 4 13 5 LS-PT 61 4 6 18 2 7 - - LS-PL 64 - - - - - 4 18 24 PT-PL - 4 10 22 0.9 9 4 25 30 # variables per class 28 2 2 4 1 2 2 4 3 Furthermore, the highest percentage of regression models with P <0.05 , was obtained by a family-level analysis for both analysis types, which increased the percentage of models from 54-65 % to 77-80 % (abundance analysis) or from 61-73 % to 73-80 % (presence- absence analysis; Table 6). However, a shift in the predictive power can be seen when comparing the two types of analyses. The presence-absence analysis typically performed better than the abundance analysis at species level, whereas the abundance analysis performed better at family level. Patch and landscape variable sets had, when grouping species to genera or families, the highest percentage of bird species that presented regression models with P < 0.05. Ill Table 6Percentage ofspecies for abundance andpresence-absence analyses (P< 0.05) for each of the different explanatory set over different taxonomic levels:family, genera, and species. Levels of taxonomic resolution Abundance analysis Presence-absence analysis Set Family Genera Species Family Genera Species LS-PT-PL 70 68 64 80 73 73 LS-PT 80 67 65 80 72 70 LS-PL 80 68 62 77 72 67 PT-PL 80 76 63 80 72 66 LS 80 67 62 77 72 63 PT 77 59 55 73 68 63 PL 80 61 54 73 68 61 Migrants and endemics Neotropical migrants analyses (Table 7), showed more models with P < 0.05 than resident bird species analyses, with 44-72 % for the abundance analysis and 52-76 % for the presence-absence analysis, whereas permanent residents showed between 40-46 % and 45- 53 %, respectively. For endemic species the percentage of models with P <0.0 5 for only meso-American endemics for abundance and presence-absence analysis was 71-81 % and 71-90 % respectively, whereas for quasi-endemics and Mexican endemics this was far less, although this is based on only 8 bird species. The analysis of species with a protection status (special protected, or threatened) produced a higher percentage of species with models with P <0.0 5 for presence-absence data than for species without protection (Table 7). The threatened species accounted for a higher percentage of models with P < 0.05 in the abundance analysis than the non-protected species. 112 ^ -R ^ k. •w R5 •2 R R <1) .0 ^t r- ? C t fr w ft,s ft. ^ 13 O s ^ 6b R a 1 ^ 1" R 1O "if -c; w s5 u R R g a; s 5 ? ^3 1 ». R -s; <£> •2 ^ K •C E-T 0 •S s R Our results supported our first hypothesis. Landscape variables explained most of the variation in bird species in both abundance and presence-absence analyses in all explanatory sets. Most previous studies suggested that landscape attributes (e.g.; composition and heterogeneity) are important factors in explaining species richness (Soderstrom and Part 2000; Wagner et al. 2000; Weibull et al. 2000). Titeux et al. (2004) had found that landscape variables explained most variation in bird species composition (27.5 %); others observed that landscape level factors explained a slightly larger amount of variation (12 %) than patch (8 %) or plot (11 %) factors (Grand and Cushman 2003). One possible reason explaining this landscape effect is that these large-scale spatial factors encompass processes and patterns (e.g.; dispersal, habitat heterogeneity) that strongly affect local species dynamics. In contrast, other studies (Miller et al. 2004; Cushman and McGarigal 2004) found that landscape variables had relatively lower explanatory power for bird community structure and composition. However, when there are strong differences among landscape elements that affect habitat quality (heterogeneity), landscape variables are likely to have greater explanatory power (Wiens etal. 1987). For the analysed bird community, the fraction of explained variance in species assemblage did not differ considerably in abundance and presence-absence coded analysis. However, the total percentage of species that presented models with P < 0.05 was higher in the presence-absence analysis than in the abundance analysis, implying that presence-absence analysis gives a better explanation of single species responses. In contrast, Cushman and McGarigal (2004) found that abundance data provided a better general measure of species- environment relationships at community level than presence-absence data. Additionally, based on our results, vegetation cover and stand structure variables, as well as the proportion of shade coffee plantations within the landscape, presented significant influence on bird species. Furthermore, human disturbances showed a significant influence on bird species. However, there is a risk of extrapolating results based on community 114 analyses to individual species. Habitat affinities for certain species may be specific to a particular region, or a set of landscapes (Miller et al. 2004). Further work is needed to determine if the different patterns observed, associated with explanatory variables at different spatial scales, are broad ecological patterns. Individual species responses are associated to one or several scales, and it is therefore necessary to repeat the analysis for individual target species. Consistent with our second hypothesis, the results demonstrated that the variation in community structure was described best at family-level than at genera- or species-level. Grouping species from species level to genera and then to family increased the percentage of species that presented models with P <0.05 . Grouping reduced the noise in responses of individual species, increasing the fit with the regression models. This is especially true for the multiple regression models and to a lesser extent to the logistic regression models. This has important consequences, as measuring richness at higher taxonomic levels, such as families, could help us to understand large-scale patterns in diversity. Previous studies suggested that using higher taxa to estimate the distribution of biological diversity retains some information on the complementarities of biotas (Williams and Gaston 1994; Williams etal. 1994;Gasto n etal. 1995). However, because there are fewer higher taxa than species, a minimum representative set of areas also tends to be smaller. Nevertheless, this is not a serious limitation because it is the more realistic case for conservation questions regarding large-scale biodiversity (Williams 1997). Additionally, higher taxa can also be used to study hotspots of endemism (Williams etal. 1994). The results of this study provide some insight into the spatial scale-dependency of diversity patterns in species-environment relationships, partitioning the effects of scale and their confounded influence on species patterns. Our study emphasizes the spatial scale- dependency of local and regional bird community structure in Cuetzalan region, Puebla, Mexico. We believe that our results give a better understanding of the relative importance of environmental factors over multiple spatially scales in explaining bird species assemblages. Furthermore, diversity patterns are dynamic at a given scale, influenced by 115 factors at broader scales (Wiens 1989). Species diversity of a local community, for example, is influenced by speciation and extinction, and by range dynamics at regional or biogeographic scales (Ricklefs 1987). Besides, individual species as well as community level analyses can be derived from this approach, which has strong implications on understanding how habitat patterns at multiple scales influence bird community structure. However, is important to emphasise that conclusions about the relative importance of explanatory variables to explain patterns are restricted by the size of the individual units of observation. What is more, a difficulty in statistical models is the assumption of causality inherent to hypotheses, because direct causal relationships may be caused by common links to other factors (Wiens 1989). Furthermore, these results could have important implications for conservation and management, specifically for bird communities of Cuetzalan region, Mexico. Conservation strategies strongly depend on aiming at the appropriate scale(s) of management, which is functionally significant for the involved species (Grand and Cushman 2003). Moreover, shade coffee plantations represent one of the main land uses in the region, and therefore by managing shade coffee plantations it is possible to provide habitat for neo-tropical migrants and forest-dependent birds, likewise endemic species or protected birds, as shown in our analysis. Acknowledgements We acknowledge the important contribution of S.Lope z de Aquino during the fieldwork phase; we also thank all members of Tosepan Titataniske Cooperative that allowed us to perform our field observations in their properties and for logistic support. We thank Idea Wild Organisation for providing essential fieldwork material. This research was supported by the CONACYT grant 149557 to E. Leyequien. 116 References American Ornithologists' Union. 1998.Check-lis t ofNort h American Birds, 7th ed: American Ornithologists' Union, Washington, D.C. Bibby C.J., BurgessN.D . and Hill D.A. 1992. Bird census techniques. Academic Press, California Bissonette, J.A.,Harrison , D.J., Hargis, CD. and Chapin, T.G. 1997.Th e influence of spatial scale and scale sensitive properties inhabita t selection by American Martins. In: Bissonette, J.A. (ed.) Wildlife and landscape ecology: effects of patterns and scale. Springer-Verlag,Ne w York, pp 310-330. Borcard, D, Legendere, P, and Drapeau, P. 1992.Partiallin g out the spatial component of ecological variation. Ecology 73(3):1045-105 5 Borcard, D. and Legendere, P. 2002.All-scal e spatial analysis of ecological data by means of principal coordinates of neighbour matrices. Ecol Modell 153:51-68 Borcard, D.,Legendere , P., Avois-Jacquet, C. and Tuomisto, H. 2004. Dissecting the spatial structure of ecological data at multiple scales. Ecology 85(7):1826-183 2 Bowman, M.F. and Bailey, R.C. 1997. Does taxonomic resolution affect the multivariate description ofth e structure of freshwater benthic macro-invertebrate communities? Can J Fish Aquat Sci 54: 1802-1807 Cushman, S.A. and McGarigal, K.2002 .Hierarchical , multiscale decomposition of species- environment relationships. Landscape Ecology 17:637-646 Cushman, S.A. and McGarigal, K. 2004. Patterns in the species-environment relationship depend on both scales and choice of response variables. Oikos 105:117-124 Dale, V.H., Pearson, S.M., Offerman, H.L. and O'Neill, R.V. 1994.Relatin g patterns of land use change to faunal biodiversity in the central Amazon. Conserv Biol 8:1027-1036 117 Dawson, D.K., Smith, D.R. and Robbins, C.S. 1997.Poin t count length and detection of forest Neotropical migrant birds. In: Ralph, C.J., Sauer, J.R. and Droege, S. (eds) Monitoring Bird populations by point counts. General technical Report PSW- GTR-149.Albany , California, pp 35-43. Gaston, K. J., Williams,P.H. , Eggleton, P.,Humphries , C.J. 1995.Larg e scale patterns of biodiversity: spatial variation in family richness. Proc R Soc Lond B 260:149-154 Gobbi, J.A. 2000. Is biodiversity-friendly coffee financially viable? An analysis of five different coffee production systems in western El Salvador. Ecol Econ 33:267-281 Grand, J. and Cushman, S.A. 2003. A multi-scale analysis of species-environment relationships:breedin g birds in apitc h pine-scrub oak (Pinus rigida-Quercus ilicifolia)community . Biol. Conserv. 112:301-317 Greenberg, R., Salgado-Ortiz, J., Warkentin, I.,Bichier , P. 1995.Manage d forest patches and the conservation of migratory birds in Chiapas,Mexico . In: Wilson, M., Sader, S. and Santana, E. (eds.) The conservation of migratory birds in Mexico. Technical Publications, University of Maine, School ofNatura l Resources,p p 178-190. ILWIS (Integrated Land and Water Information System) 1997.Applicatio n and reference guides. ILWIS DepartmentJTC, Enschede, TheNetherlands , p. 352 Jacobsen, D. 2004. Contrasting patterns in local and zonal family richness of stream invertebrates along an Andean altitudinal gradient. Freshwater Biol. 49:1293-1305 Kent, M. and Coker, P. 1992. Vegetation description and analysis. Apractica l approach. John Wiley & Sons,Ne w York. Kneitel, J.M. and Chase, J.M. 2004. Trade-offs in community ecology: linking spatial scales and species coexistence. Ecol. Letters 7:69-80 Levin, S.A. 1992.Th e problem of pattern and scale in ecology. Ecology 73:1943-1967 118 Levin, S.A. 1993.Concept s of scale at the local level. In: Ehleringer, J.R. and Field, C.B. (eds.) Scaling physiological processes: leaf to globe.Academi c Press,Ne w York, New York, USA. Liang, L., Stocking, M., Brookfield, H. and Jansky, L. 2001. Biodiversity conservation through agrodiversity. Global Environmental Change. 11:97-101 Lillesand, T.M. and Kiefer, R.W. 1979. Remote Sensing and Image Interpretation. John Wiley & Sons,Ne w York. Liu, Q.H. and Brakenhielm, S. 1995.A statistical approach to decompose ecological variation. Water, Air, and Soil Poll.1-4:61-8 7 Loreau, M. 2000. Are communities saturated? Onth e relationship between a, Pan dy diversity. Ecol.Letter s 3:73-76 Meentemeyer, V. and Box, E.O. 1987. Scale effects in landscape studies. In: M. G. Turner (ed.)Landscap e heterogeneity and disturbance. Springer- Verlag, New York, USA,p p 15-34 Miller, J.R., Dixon, M.D. and Turner, M.G. 2004.Respons e of avian communities in larger-river floodplains to environmental variation at multiple scales. Ecological applications 14(5):1394-141 0 Moguel, P. and Toledo,V.M . 1998.Biodiversit y conservation in traditional coffee systems of Mexico. Conserv. Biol.13(1) :11-2 1 Mouquet, N., Munguia, P., Kneitel, J.M. and Miller, T.E. 2003. Community assembly time and the relationship between local and regional species richness. Oikos 103:618- 626 O'Neill, R.V., Hunsaker, C.T., Jones, K.B.,Ritters ,K. , Wickham,J.D. , Schwartz, P.M., Goodman, I.A., Jackson, B.L. and Baillargeon, W.S. 1997. Monitoring environmental quality at the landscape scale.Bioscienc e 47:513-519 Parker, T.A. 1991. On the use of tape recorders in avifaunal surveys. TheAU K 108(2):443- 444 119 Ricklefs, R.E. 1987. Community diversity: relative roles of local and regional processes. Science 235:167-171 Soderstrom, B.an d Part, T.2000 . Influence of Landscape Scale on Farmland Birds Breeding in Semi-Natural Pastures. Conserv. Biol. 14(2):522-533 Soto-Pinto, L.,Perfecto , I., Castillo-Hernandez, J. and Caballero-Nieto,J . 2000. Shade effect on coffee production at the northern Tzeltal Zone of the state of Chiapas, Mexico.Agric , Ecosyst. and Environ. 80:61-69 Titeux,N. , Dufrene, M., Jacob, J.P.,Paquay , M. and Defourny, P. 2004. Multivariate analysis of a fine-scale breeding bird atlas using ageographica l information system and partial canonical correspondence analysis: environmental and spatial effects. J. Biogeogr. 31:1841-1856 Van Langevelde, F.2000 . Scale habitat connectivity and colonization in fragmented nuthatch populations. Ecography 23:614-622 Vielliard, J.M. 2000. Bird community as an indicator ofbiodiversity : results from quantitative surveys. Brazil. A. Acad. Bras. Ci. 72(3):323-330 Wagner, H.H., Wildi, O., Ewald, K.C. 2000.Additiv e partitioning ofplan t species diversity in an agricultural mosaic landscape. Landscape Ecol. 15(3):219-227 Weibull, A.C.,Bengtsson , J. andNohlgren, E. 2000.Diversit y of butterflies in the agricultural landscape:th e role of farming system and landscape heterogeneity. Ecography 23:743-750 Wiens,J.A. , Rotenberry, J.T.an d Van Home, B. 1987.Habita t occupancy patterns ofNort h American shrub steppe birds:th e effects of spatial scale. Oikos 48:132-147 Wiens, J.A. 1989. Spatial scaling in ecology. Func. Ecol. 3:385-397 Williams. P.H. and Gaston, K.J. 1994.Measurin g more ofbiodiversity : can higher-taxon richness predict wholesale species richness? Biol. Conserv. 67:211-217 120 Williams,P.H. , Humphries, C.J., Gaston, K.J. 1994.Centre s of seed-plant diversity: the family way. Proc. R. Soc.Lond . B 256:67-70 Williams, P.H., Gaston, K.J. and Humphries, C.J. 1997.Mappin g biodiversity value worldwide: combining higher-taxon richness from different groups. Proc. R. Soc. Lond. 264:141-148 121 Chapter 6 Capturing the fugitive: applying remote sensingt o terrestrial animaldistributio n anddiversity .A review. Euridice Leyequien, Jochem Verrelst, Martijn Slot, GabrielaSchaepman-Strub, Ignas M.A. Heitkonig,Andrew Skidmore Abstract Amongst many ongoing initiatives to preserve biodiversity, the Millennium Ecosystem Assessment again shows the importance to slow down the loss of biological diversity. However, there is still a gap in the overview of global patterns of species distributions. This paper reviews how remote sensing has been used to assess terrestrial faunal diversity, with emphasis on proxies and methodologies, while exploring prospective challenges for the conservation and sustainable use of biodiversity. We grouped and discussed papers dealing with the faunal taxa mammals, birds, reptiles, amphibians, and invertebrates into five classes of surrogates of animal diversity: 1. habitat suitability, 2.photosyntheti c productivity, 3. multi-temporal patterns, 4. structural properties of habitat, and 5. forage quality. It is concluded that the most promising approach for the assessment, monitoring, prediction, and conservation of faunal diversity appears to be the synergy of remote sensing products and auxiliary data with ecological biodiversity models,an d a subsequent validation of the results using traditional observation techniques. 124 Introduction The importance ofbiodiversit y conservation iswidel y recognised asther e is a general concern about its current status and about the responses by society topresen t and future environmental changes (Gaston, 2000; Mace, 2005,Millenniu m Ecosystem Assessment, 2005). Biodiversity definitions include different levels of organisation of biological variation and richness, from genes and species to ecosystems. Noss (1990) expressed the variation at each ofthes e hierarchical levels in terms ofthre e spheres: composition (e.g. the genes of different cattle races), structure (e.g.th e ratio of large versus small bodied animals), and function (e.g. forage consumption). However, this multi-faceted nature of the term biodiversity makes it a difficult concept to capture in one definition or description, so it cannot be measured in a single parameter (Noss, 1990;Wolfang , 2003; Scholes and Biggs, 2005). The challenge to measure these levels and spheres of organisation of biological variation has led to the search ofrelevan t biological indicators from which biodiversity could be measured. These indicators include species, habitats, and eco-regional characteristics, which can be sampled in the field, categorized, and interpreted. Despite the efforts of scientists and policy makerst o reduce the rate of species loss,ther e is still a gap in the overview of continental and global patterns of species distributions (Brooks etal, 2001; Ceballos et al, 2005). Remotely sensed data contribute to the assessment and monitoring of biodiversity from local to global scales (Murthy etal, 2003), and over time, with spatially continuous coverage. Since the 1980s, satellite multispectral imagery became a common tool,particularl y in exploring the composition of biodiversity, i.e., species richness (e.g. Saxon, 1983;Nagendra , 2001). Several articles reviewed the potential and contribution of remote sensing data products to assess terrestrial vascular plant species diversity (e.g., Stohlgren etal, 1997;Goul d 2000; Griffiths et al, 2000; Nagendra 2001). This review article has the purpose to summarise the historical development and prospective approaches in which remote sensing was used to assess and monitor terrestrial faunal diversity. While an important methodology ofplan t diversity consists in direct mapping of species and associations (Nagendra 2001), the fugitive and secretive nature of animals requires approaches based on proxies and surrogates. Based on 125 the currently applied methodologies, the following broad categories were identified, reflecting chronological approaches inthi s field: 1.habita t suitability mapping, relying on species-habitat associations, 2. spatial heterogeneity assessment based on primary productivity, 3.tempora l heterogeneity assessment, 4. mapping of structural properties of habitat, and 5.mappin g ofplan t chemical attractants, relying ofth e influence of land cover attractants on fauna, such as forage quality. These latter categories enclose the range from the most frequently used methodologies applied to terrestrial animal taxa to the latest approaches found in literature. In addition, direct and indirect measurements of species diversity and distribution are illustrated. Within each approach, the literature was furthermore grouped following the taxonomical system: mammals, avifauna, reptiles, amphibians, and invertebrates. These taxonomic groups represent the most frequently studied taxa, and reflect different challenges to the application of remote sensing to assessing animal species presence. The discussion evaluates theory development and the potential use ofremot e sensing techniques for terrestrial animal ecology studies related to species diversity, and the prospective direction ofremot e sensing approaches applied to this field. Assessing speciesrichnes sthroug h habitat suitability mapping The most straightforward approach to estimate animal distribution or species richness from remotely sensed data ist o identify and detect animal habitat suitability. A habitat isth e local environment in which an organism normally lives and grows.I n order to map habitat, knowledge ofhabita t preferences and the requirements ofth e species of interest is combined with airborne or satellite data, biophysical, geophysical data, and meteorological data. For ecological biodiversity assessments, field surveys areusuall y executed to collect data on species distribution, habitat use or characteristics of nesting, breeding orburrowin g sites. Additionally, habitat usepattern s can be derived from analysis of movements ofradi o or satellite collared individuals (e.g. Kanai etal, 1994;Bechte l et al, 2004). Using remote 126 sensing, these local measurements can be extrapolated to cover a large region of interest, and estimate habitat suitability. After collecting field survey data, spectral insitu measurements at the locations ofth e ecological assessments or the spectral properties of the pixel corresponding to this location are used astrainin g data to classify the imagery for a larger area. Resulting maps with spatially discrete habitat types can then be analysed using a wide array of statistical techniques to validate classified habitat with species population data. The main problem with this approach isth e assumption that empirical conditions at the field survey point may be extrapolated over a large area. Such an assumption needs to be carefully tested otherwise the resulting maps will be biased by the sample points. In other words, habitats may not be described and stratified in ecologically meaningful terms, which could limit the predictive value of the relationships between reflectance data and species distribution within and beyond the study area. The following sections demonstrate how widely remote sensing approaches are applied to estimating habitat suitability in terrestrial environments throughout many animal taxa. Mammals One ofth e earliest publication involving satellite imagery to detect mammalian species dates from 1980,whe n Loffler and Margules estimated the distribution of hairy nose wombat (Lasiorhinus latifrons) in southern Australia by identifying burrows and mounts from Landsat imagery. Also inAustralia, Saxon (1983) used Landsat imagery to locate a habitat suitable for the re-introduction of rufous hare-wallabies Lagorchestes hirsutus. Later in the 1980s this approach ofrelatin g remotely sensed land cover types to habitat suitability, was used amongst others for assessment of the habitat ofth e giant panda, Ailuropoda melanoleuca (De Wulff etal, 1988),elk , Cervus elaphus (Eby and Bright, 1985) and white-tailed deer, Odocoileus virginianus borealis, (Ormsby and Lunetta, 1987). The method of habitat mapping by means of signature classification was applied throughout the 1990s (e.g. Huber and Casler, 1990;De l Valle etal, 1997;Fulle r et al, 1998; Cardillo etal., 1999;Richards , 1999) and is still used today (e.g. Oindo etal, 2003; Sharma et al, 2004). 127 Other authors tried to derive general patterns of species richness in relation to habitat (White etal, 1997;Maso n etal, 2003).Fulle r etal. (1998 ) combined field surveys of plants and animals with satellite remote sensing ofbroad vegetation types to map biodiversity inth e Sango Bay area in Uganda. They identified 14land-cove r classes from reflectance characteristics and validated the results with field surveys,recordin g 86% correspondence between field and map data. The field surveys included flowering plant species, dragonflies, butterflies, fish, amphibians, reptiles,bird s and mammals. These species data were used to generate biodiversity ratings,base d on species 'richness' and 'rarity', which could be related to the vegetation cover. Similarly, Cardillo etal. (1999 ) predicted species richness and occurrences ofterrestria l mammals from Principal Component Analysis (PCA) ordinated land cover variables from the Land Cover Map of Great Britain. Because of the high proportion of species with geographic distributions changing independently of land cover, the predictive strength ofth e land cover data for species richness assessment was however limited. Heitkfinig etal. (2003 ) directly correlated the distribution of large mammalian herbivores inth e Okavango delta in Botswana with Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM+) reflectance data. Animal distribution was sampled along field transects,whil e registering the locations with a Global Positioning System (GPS), and the recorded animal presence was related to the spectral signature of the location. A multivariate approach allowed for the distinction of occurrence of several mammalian species, including elephant (Loxodontaafricana), zebra (Equus burchelli), and impala (Aepyceros melampus). However, species with low densities in the field, including giraffe (Giraffa camelopardis)an d wildebeest {Connochaetestaurines), were not successfully distinguished. Bechtel etal. (2004 ) used a comparable method for woodland caribou {Rangifertarandus caribou). Because habitat maps were incomplete for their study area, they correlated spectral information obtained from Landsat 5T M satellite data with Global Positioning System (GPS) locations of satellite collared, and Very High Frequency Radio locations of collared animals. First, a statistical approach was used to automate the classification of each satellite imagery pixel to generate landscape classes based on 128 characteristic spectral signatures. These classes were then regrouped corresponding to their relation with the GPS-registered animal presence. The resulting map indicated the woodland caribou use and avoidance of areas,base d on corresponding identified spectral classes ofth e satellite imagery. Coops and Catling (1997)use d airborne multispectral videographic data to accurately predict the complexity of fauna habitat across forested landscapes. This video system provides atool for stratifying the forest into fauna habitats topredic t the composition, spatial distribution and abundance of faunal groups that are known toprefe r the eucalypt forests (Coops et al, 1998;Catlin g et al, 2000). Coops and Catling (2002) then related habitat quality based on complexity scores, topredic t future relative abundance ofth e long- nosed potoro {Potoroustridactylus) and the large wallabies (red-necked wallaby, Macropus rofogriseus, and swamp wallaby, Wallabiabicolour) acros s landscapes. Mammals are relatively well studied and their habitat preference isquit e well documented, which iso f vital importance for successfully correlating mammal occurrence to remotely sensed habitat data. Many species however, (e.g. generalist species) use more than a single distinct vegetation type, and non-herbivore species tend to have little direct association with a habitat or vegetation type that can be remotely sensed (e.g. Cardillo etal, 1999; Cowley etal, 2000). In other cases,ther e is a limitation to the use of remotely sensed data due to the animals' elusive nature. Thepredictiv e value of mammal-habitat relationships will also be limited for species for which the habitat preference changes with geographical position (e.g., Cardillo et al, 1999).Anothe r issue that complicates the link between species presence and habitat characteristics isth e socio-biology of the species. The predicted distribution of guanaco (Lamaguanicoe), for example,prove d to have little correlation with the real distribution and densities (Del Valle etal, 1997). The socio-biology of the species, classified as 'resource defence polygyny' (Franklin, 1983)wher e a dominant male will defend his territory against other males, in combination with interspecific competition with sheep and anthropogenic influences, was identified asth e possible reason for the inaccuracy ofth e predictions. 129 Avifauna Since the 1980s,remot e sensing has been widely used in assessing and monitoring bird distribution and habitats (e.g. Baines et ah, 1986;Perra s etal, 1988,Minto n etah, 2003; Venier etah, 2004).Habita t maps are aggregated from land cover maps that are generally produced from Landsat or radar imagery. The habitat maps are then combined with data frombir d composition and abundance surveys, yielding distribution and density maps (e.g. Palmeirim, 1988; Schwaller etah, 1989;Aver y and Haines-Young, 1990; Kanai et ah, 1994; Morrison, 1997;Debinsk i etah, 1999;Osborn e etah, 2001; Taft etah, 2003; Fuller etah, 2005;Prin s etah, 2005). Apart from Miller and Conroy (1990),wh o classified data from the Satellite Probatoire d'Observation de la Terre (SPOT) topredic t Kirtland's warbler {Dendroica kirtlandii) occurrence inth e Bahamas based on vegetation types, most studies applied Landsat TM data for the prediction of single bird species habitat suitability. Knowledge and ground studies on habitat preferences for nesting and feeding were combined with Landsat data for sedentary (e.g. Osborne et ah, 2001; Johnson etah, 1998;Hurlber t and Haskell, 2003) as well as migratory birds (e.g. Green etah, 1987; Avery and Haines-Young, 1990; Sader et ah, 1991). Environmental criteria (i.e.vegetatio n cover, landscape characteristics) derived from Landsat data were used to assess the location of nesting sites ofbuzzards , Buteo buteo, (Austin etah, 1996) and great sandhill cranes, Gruscanadensis, in Minnesota (Herr and Queen, 1993).Further , the probability of occurrence of 14ou t of2 3 species of land birds in Maine (Hepinstall and Sader, 1997), as well as winter distributions of sage grouse (Centrocercus urophasianus) in Utah (Homer etah, 1993) were successfully predicted using Landsat imagery. Kanai etal. (1994 ) analysed data of satellite collared red crowned cranes (Grusjaponensis), hooded cranes (Grus monacha) and white-naped cranes (Grus vipio) to determine habitat use and species distribution. Subsequently, a combination of Landsat with Marine Observation Satellite- Multi-spectral Electronic Self-scanning Radiometer (MOS1 MESSR) data was used to derive the characteristics of seven sites that 130 were identified as crane habitat. The method proved very helpful, because the large crane habitats cannot easily be monitored through ground studies. Some recent studies focus on avifaunal diversity, leading from single species assessments to diversity predictions for whole taxa. N0hr and Jorgensen (1997)relate d avian diversity in northern Senegal to landscape diversity, while Debinski etal, (1999) categorized habitats in the Greater Yellowstone ecosystem based on Landsat data and then determined the relationship between habitat categorizations and plant, bird, and butterfly species distribution patterns. They found that sites ofhighes t species richness coincided for plants, birds, and butterflies. This coincidence of'hotspots' of different taxa shows those indicator taxa could beuse d to assess an areas' biodiversity status.Als o inth e Greater Yellowstone ecosystem Saveraid etal. (2001 ) assessed potential bird habitats for 11type s of montane meadows. Landsat imagery was further analyzed to identify habitats for migratory birds in Costa Rica (Sader etal, 1991), andt o predict bird species richness inborea l agricultural- forest mosaics in south-western Finland (Luoto etal, 2004). Fuller etal (1998) related the diversity of various taxa, including flowering plant species, dragonflies, butterflies, fish, amphibians, reptiles,bird s and mammals in Uganda to vegetation cover derived from Landsat. Areas of bird endemism in East Africa were related to remotely sensed climatic variables by Johnson etal. (1998) ,t o obtain general patterns ofbir d species richness. It was concluded that contemporary environmental conditions,ultimatel y determined by climate, appear to account for a substantial fraction ofth e observed variation inth e distribution of endemic bird areas. The above bird studies demonstrate that habitat classifications based on remote sensing data can be successful for sedentary and migratory birds, aswel l as bird communities. However, the spatial and spectral scale ofth e data appears to be a crucial factor in the prediction of bird occurrence patterns. Laurent etal. (2005 ) investigated thepotential of using unclassified spectral data for predicting the distribution of three bird species by varying 1) the window sizeuse d to average spectral values in signature creation, and 2)th e threshold distance for recording bird observations. Accuracy statistics for each species were affected 131 by the detection distance ofpoin t count surveys used to stratify plots intopresenc e and absence classes. Thus, the accuracy of wildlife occurrence maps classified from spectral datawill differ given the species of interest, the spatial precision of occurrence records used asgroun d references and the number ofpixel s included in spectral signatures. Akin to mammal studies, a low number ofbir d records per habitat type decrease the power of the statistical tests to distinguish differences between habitat use and availability in bird studies (Garshelis 2000). Reptiles and amphibians Onlytw o studies on mapping reptilian or amphibian habitats using remote sensing data were identified, but both were innovative and successful. Raxworthy etal. (2003 ) assessed and predicted the distribution of known and unknown chameleon species in Madagascar, using a combination of satellite data (Moderate Resolution Imaging Spectroradiometer (MODIS) and historical and recent chameleon observations on the island. A generic algorithm for rule-set prediction (GARP) was used to delineate ecological niches, based on environmental geographical information system (GIS) data, and to predict geographical distributions of species. This study leads to the discovery of seven new species of chameleon. Scribner et al.(2001 ) used in situ and remotely sensed data of the aquatic and terrestrial environment, to examine the correlation ofhabita t characteristics with population demographic and genetic characteristics of the common toad (Bufo bufo). This study was the only one (encountered) that focussed onth e sub-species level.Alleli c (i.e. genetic) richness,populatio n size, and toad presence were mostly associated with terrestrial habitat variables, likepon d density, availability ofwoodlands , hedgerows, and anthropogenic development. 132 While remotely sensed data for animal diversity assessment using habitat characteristics is increasingly used, its application to reptile and amphibian diversity remains poorly explored. Despite the success ofth e above presented studies,ther e is still aga p between ecological theory and the application ofremotel y sensed data. One of the problems is that there isn o clear understanding over which spatial scales the species-habitat relationships apply for species of interest, specifically those of limited vagility. The complex life histories of amphibians and their secretive behaviour add to the challenge of successfully using remotely sensed data. Invertebrates Remotely sensed imagery is increasingly used to detect insect habitats or the effects of insects on their environment (Riley, 1989;Hay , 1997). Habitat patch characteristics essential for many insects, such asmicro-variation s inwetlands , grasslands and forests are too small to be identified using conventional imagery (Fisher, 1997;Cracknell , 1998).A recent study tackled the problem of sampling scale inherent to insect habitat mapping by comparing two satellite sensorswit h different spatial resolution (SPOT and Landsat) and, as a result, optimized insect species richness mapping at a landscape level (Chust et ah, 2004). One ofth e most common cases of insect habitat mapping deals with swarming insects that have the potential to destroy their habitat. Defoliating insects, oncepresen t in a food- endowed environment, quickly expand to adevastatin g outbreak consuming their primary resource. This allows aprecis e distribution mapping through classification of vegetation defoliation or discoloration using remote sensing. For instance, Joria et al.(1991 ) successfully classified gypsy moth (Choristoneurafumiferand) defoliation into three damage classes by delineating affected areas with Landsat data. Franklin and Baske (1994) differentiated four defoliation levels of spruce budworm (Choristoneurafumiferand) in a balsam fir forest analyzing SPOT HRV data. The spectral response of the Jack Pine (Pinus banksiana) canopy, attacked by the Jack Pine budworm (Choristoneurapinus pinus), was 133 first described by Hall etal. (1995) . Landsat TM images before and after defoliation were acquired to mapth e top-kill severity. Another aggressive defoliator actor widely documented isth e mountain pine beetle (Dentroctonusponderosd) which is primarily hosted by the lodgepole pine (Pinus contorta). Combinations of SPOT multispectral and panchromatic bands and PCA-transformations were evaluated by Sirois and Ahern (1988) to determine their ability to detect mountain pine beetle mortality ('red attack'). A conceptual model based on spectral brightness-greenness was developed and tested (Price and Jakubauskas, 1998) to relate beetle infestation to spectral properties. Similar examples are reported for pear thrips (Taeniothrips inconsequens) (Vogelmann and Rock, 1989);th e black-headed budworm (Aclerisvariana) (Franklin etal, 1994;Luthe r etal, 1997);th e aspen tortrix (Choristoneuraconflictana) (Hall et al, 2003);an d the Douglas-fir beetle {Dendroctonuspseudotsugae) (Lawrence and Labus,2003) . Inmor e diverse ecosystems, notable efforts to create habitat maps from remotely sensed data were made for butterfly and beetle species. Butterfly species are often host-specific and their diversity may correlate with underlying plant diversity. Thus, Debinski et al (1999)reporte d that several rarebutterfl y species significantly correlated with remotely sensed habitat types in the Greater Yellowstone Ecosystem. The modelling of Luoto et al (2002) supported the findings that specialist butterfly species distribution is closely related to remotely sensed habitat types. British ground (Coleoptera, Carabidae) (Eyre et al, 2003a) and water beetle {Coleopteraspec.) species pool distribution (Eyre etal, 2003b) strongly correlated with satellite-derived land cover data. Chust et al.(2004 ) successfully assessed woodland invertebrate taxa distribution based on satellite imagery. Compared to othertaxonomi c groups,th e ability of remote sensing data to contribute to the mapping andpredictio n of occurrence of invertebrate diversity appears tob e poorly investigated. The current trend of studies onterrestria l invertebrates shows an emphasis on insects,wherea s for other taxa there isn o literature known to the authors. The majority of articles concentrated on insects that are considered pests, and their effects on crops or forests. However, very few studies deal with conservation efforts. Besides, the approach is 134 limited to habitat-insect relationships using only beta diversity (i.e. local diversity, within- community component) whereas gamma diversity (i.e.tota l regional diversity) is not taken into account. Summary Habitat suitability iswidel y used as aremotel y sensed proxy for species distribution and richness. It mainly covers the composition sphere of biodiversity. Though successful in many ofth e discussed examples, the micro-heterogeneity of an area required for many species does not always allow a discrete classification approach. Many species (e.g. generalist species) use more than a single distinct vegetation type and some non-herbivore species may show low strength of association with ahabita t or vegetation type because many species, regardless ofth e degree of habitat-specificity, do not occupy the full extent of their preferred habitat type that can be remotely sensed (e.g. Cardillo etal, 1999; Cowley etal, 2000). Current habitat classification isbase d on discrete maps and the resulting representation of class boundaries may not capture the meaningful ecological functional variability for each species. Correspondence between field data and remotely sensed imagery aimed at species communities was found to behig h in some studies (Fuller etal, 1998;Bechte l et al, 2004),bu t limited in others (e.g. Cardillo etal, 1999;HeitkOni get al, 2003). One factor limiting the accuracy in this approach appears tob e the application ofproxie s at inappropriate spatial, spectral, and temporal resolutions. Remote sensing studies involving species diversity need to consider different levels of taxonomic resolution. Several studies used ahighe r or lower taxonomic resolution approach asprox y for estimating species richness for other taxa (Baldi, 2003; Olsgard etal, 2003; Doerries and Van Dover, 2003; Sauberer etal, 2004; Ward and Lariviere, 2004). Cross-taxon congruence in biodiversity across different groups of organisms was also investigated aspotential surrogates for each other (Negi and Gadgil,2002 ;Hein o etal, 2005). However, correlations and congruencies in species richness among different taxonomic groups are difficult to generalise as they 135 differ to environmental gradients.Accurac y of assessing species diversity in particular may further increase by adding environmental variables to the analysis. Moreover, despite the potential of remotely sensed data for habitat suitability analysis, ground survey data (e.g., species composition, abundance, and density maps) are essential to provide the basis for finding ecologically meaningful interpretations and for predicting species distribution and diversity. Assessing species richness through spatial heterogeneity based on primary productivity Spatial heterogeneity is one ofth e driving factors in the explanation of species richness (Stoms and Estes, 1993). Itha s long been accepted that environmental heterogeneity may support richer species assemblages compared to simple ecosystems (Simpson, 1949; MacArthur and Wilson, 1967;Lack , 1969;Huston , 1994)becaus e of the creation of niche differentiation (Tilman etal., 1997;Loreau , 1998). This iso f particular relevance for dealing with the structural sphere ofbiodiversity . Here, we are concerned not only with the species composition, but also with the relationships of species towards one another. Factors contributing to the environmental heterogeneity include the temporal and spatial variation in the biological, physical, and chemical features of the environment that create different conditions that species can preferentially exploit (Morin, 2000).I n comparison with the previously discussed discrete classification approach, the biological, physical, and chemical features are represented in a continuous way. Depending on the spatial, spectral, temporal, and angular resolution of the remotely sensed data, different levels of differentiation are reached, while post-processing techniques (e.g. density slicing, thresholds) allow the assignment of a discrete class to every pixel, ifnecessary . Plant productivity and biomass of ecosystems vary in space and time, andth e spatial heterogeneity in productivity is hypothesized to influence species distribution and local abundance of individuals (Brown, 1988; Currie, 1991; Brown and Lomolino, 1998;Gasto n and Blackburn, 2000; Oindo and Skidmore, 2002; Seto,2004) . 136 The most commonly used parameter for quantifying productivity and above-ground biomass of ecosystems isth e Normalized Difference Vegetation Index (NDVI) (Tucker, 1979).I t is based onth e strong absorption ofth e incident radiation by chlorophyll in the red, and the contrasting high reflectance byplan t cells inth e Near infrared (NIR) spectral region. Because it isbase d on the normalized ratio of the reflectance in these two spectral bands (i.e.,NDV I = (NIR-red)/(NIR+red)), it is an indicator ofth e greenness of vegetation canopies and able to separate vegetation from other materials.NDV I values proved to be a suitable indicator for vegetation parameters including biomass and above-ground primary productivity (e.g., Sellers, 1985, 1987;Tucke r and Sellers, 1986;Bo x etal, 1989), and iti s therefore often correlated to faunal species occurrence and diversity. Mammals Since the late 1990s, an increasing number of studies is analysing NDVI to predict wildlife habitat suitability. Verlinden and Masogo (1997) found a significant positive relationship between NDVI and grass greenness inth e Kalahari of Botswana. The relationship between NDVI and animal distribution using animal census data turned out tob e more complex. Results using presence/absence data indicated a significant selection for higher NDVI signatures only for ostrich {Struthiocamelus) and wildebeest {Connochaetestaurinus), the latter onlywhe n present in high numbers. The gemsbok {Oryxgazelle), the less abundant eland {Taurotragusoryx), and the locally concentrated springbok {Antidorcas masupialis) did not show significant relationships with greenness andNDVI . Musiega and Kazadi (2004) found, that the great seasonal migration ofherd s ofwildebees t {Connochaetes taurinus) in the Serengeti-Mara ecosystem isprimaril y driven by green vegetation availability, as detected usingNDVI . In another African case, Zinner etal. (2001 ) described habitat quality in central Eritrea through NDVI derived from Landsat MSS satellite data for three baboon species {{Papio hamadryas hamadryas, Papio hamadryas anubis and Chlorocebus aethiops). Hamadryas {Papioh. hamadryas) and olive baboons {Papioh. anubis) tended to select better quality habitats, characterized by ahighe r NDVI than the average in four out of five ecogeographical zones in Eritrea. Moreover, Hamadryas 137 baboons showed agreate r ecological plasticity than olive baboons,whic h are confined to riverbeds with extended gallery forest. Although successful in some studies (e.g. Zinner et al., 2001), difficulties in correlating NDVIwit h the mammal distribution of less abundant species (Verlinden and Masogo, 1997)remai n unsolved. Results suggest that relationships between less abundant species and greenness might become insignificant because of a large number of unoccupied suitable habitats. Moreover, the biomass-based approach is successful only with herbivorous species that are sensitive to differences in vegetation characteristics across an area. Avifauna Jergensen andN0hr (1996) and Nehr and Jorgensen (1997) used a combination of satellite image analysis and ornithological surveys to assess avian biodiversity in the Sahel.A Landsat image was used to derive landscape diversity andNDVI , the latter being an indicator for the annual biomass production. Both variables were significant factors ina multiple regression model explaining species diversity. Hurlbert and Haskel (2003) analysed avian species richness inrelatio n to primary productivity and habitat heterogeneity inAmerica . They found thatNDV I was a good predictor of seasonal species richness at fine spatial scales,wherea s habitat heterogeneity best predicted richness at coarser spatial resolutions. Hawkins etal. (2003,2004 ) showed that productivity indicators (NDVI and actual evapotranspiration) correlated well with bird diversity data inNort h America. A positive correlation betweenNDV I and bird and butterfly species richness was found by Seto (2004),thoug h this relation did not have a definite functional shape. The relationship between NDVI and species richness ofbutterflie s was strongest athig h spatial resolutions, whilst that ofbird s was better at a lower resolution. In a comparable study, Bailey et al. (2004) distinguished between both habitat primary productivity and habitat heterogeneity by using estimated maximum NDVI and the spatial variation therein. This, in turn, was correlated with species richness of birds and butterflies. They found positive linear relationships between maximum NDVI and the number of functional guilds of birds and species richness of neotropical migrant birds,bu t a negative association betweenNDV I and 138 the number of functional guilds of birds and species richness ofresident birds. Alternatively, Lee etal. (2004 ) found a hump-shaped relationship between NDVI and bird species richness in Taiwan, but this became insignificant when effects of roads and elevation were accounted for. Generally, NDVI proves to be a suitable proxy reflecting primary production or heterogeneity. Nevertheless, the correlations with bird species diversity were positive, hump-shaped, or even negative. The differences in results suggest that a functional link between NDVI and diversity remains elusive, underpinning the importance of ground truth data and validation. Results from several studies (e.g.,Baile y etat, 2004;Cushma n and McGarigal, 2004) suggest that taxa related scale issues aret o be considered when setting up a monitoring scheme using remote sensing. Reptiles and Amphibians In a study on two genetically differentiated forms of the Golden-striped salamander {Chioglossa lusitanicd) in Portugal, Arntzen andAlexandrin o (2004) applied GIS-based rules in addition toNDV I data, and found that the southern form of the salamander tended to encounter harsher environmental conditions, with lower precipitation, air humidity, summer temperatures andNDVI , but with a higher number of frost months than the northern form. This isth e only study on amphibians usingNDV I to assess or monitor species richness. Invertebrates Very few studies use remote sensing data to assess orpredic t invertebrate species richness beside cases cited above aspest s or acting asdiseas e vectors.Nevertheless , Seto et al. (2004) and Bailey etal. (2004 ) found strong correlations between NDVI values and butterfly species' richness inth e Great Basin ofwester n North America. 139 Insect outbreaks may result in such a dramatic reduction in standing biomass, that it enables vegetation indices to precisely indicate the affected location. Amongst the earliest reports on remote insect detection, Nelson (1983) analysed Landsat data for detecting significant forest canopy alteration caused by gipsy moth (Lymantriadispar) defoliation. The author found that the transformed vegetative index difference (VID= NIR - red) most accurately delineated forest change, andthu s was able to map gypsy moth outbreak. More recently, MODISNDV I data was successfully analysed to map a locust {Locustmigratoria manilensis) plague in China (Ma etal, 2005). TheNDV I difference image between the data before and after the peak damage ofth e locust plague accurately mapped the geographical extent and severity of the affected areas. Many invertebrate studies used NDVI in combination with the reflectance in the middle infrared, land surface temperature, and rainfall to predict abundance, distribution and seasonality of diseases transmitted by an invertebrate vector. Robinson etal. (1997 ) used satellite data to make predictions of the probable distribution oftsets e fly (Glossina spp.) species in southern Africa. For some subspecies (e.g. Glossina morsitans centralis) the distribution was best correlated with NDVI and the average maximum temperature(75 % correct predictions). Relative abundances of the midge Culicoides imicola,th e vector of bluetongue virus and African horse sickness virus, atvariou s sites in Morocco and Spain, were compared with climatic variables, altitude andNDV I ofth e same sites (Baylis et al, 1998).N o significant correlations were found, although wind speed and NDVImin explained over 50%o f the variance in abundance. Using broadband NOAA AVHRR data, Hay etal. (1998 ) successfully correlated NDVI-series with malaria presence in Kenya, and malaria admissions could be predicted across Kenya in an average year with regression analysis.NOA A AVHRR data of a site in Brazil (Bavia etal, 2001) and sub-Saharan Africa (Kristensen etal, 2001; Malone etal, 2001) was similarly used to produce NDVI maps, and analyzed for relationships with the prevalence of schistosomiasis, hosted by snails. Results indicated thatNDVI , together with climate data,predicte d snail distributions accurately enough for schistosomiasis risk assessment. 140 As in the case of habitat characteristics used topredic t species distribution, the use of NDVI has a strong emphasis on insects considered apes t and those that act as vector diseases,thu s there is a lack of conservation-oriented studies for species which are not considered aspests . The use ofNDV I has proven to be successful, however it depends directly on the species life history and ecology whether NDVI can act as a surrogate itself or in combination to other remotely sensed data. Summary The approach of assessing species distribution and richness through spatial heterogeneity based on primary production, can be considered as a functional, non-discrete correlation. Here, animal occurrence and diversity are related to terrestrial features by means of an ecological link. The link emphasised in this review is atrophi c one (i.e.food-related) , e.g. the case of herbivore animals being correlated with local vegetation biomass or primary productivity (Oindo,2002 ; Seto etal, 2004), and heterogeneity therein. The heterogeneity hypothesis affirms apositiv e relationship between ecosystems diversity and biological diversity (Simpson, 1949;MacArthu r and Wilson, 1967;Lack , 1969;Hutson , 1994).Th e spatial and temporal heterogeneity inprimar y productivity is an explanatory variable to assess species occurrence and richness. Although successful in some studies (Zinner et al, 2001; Ito etal., 2005) , difficulties to correlate NDVI and animal distribution of less abundant mammal species (Verlinden and Masogo, 1997)remain s unsolved. Moreover, the biomass-based approach is successful only with species sensitive to differences across an area. More elaborate studies, including additional explanatory environmental variables together with primary productivity heterogeneity (e.g. landscape diversity, evapotranspiration, land surface temperature, rainfall, altitude), explained considerable variation in species richness,bu t -her e too -th e explanatory power of each variable differed among spatial scales (Robinson et al, 1997; Baylis et al. 1998;Hurlber t and Haske, 2003; Hawkins etal, 2003, 2004;Baile y et al 2004). 141 One ofth e factors influencing the accuracy ofprediction s of species richness using primary productivity indicators (NDVI) is scale orresolution , where the variation of species diversity -withi n and between taxa - isbette r explained using a specific scale, e.g. finer spatial resolution, whereby other variables explains more ofth e variation at another scale, e.g. coarser spatial resolution (Bailey etal, 2004; Cushman and McGarigal, 2004; Hawkins etal, 2003; Lee et al, 2004;Hurlber t and Haskel, 2003). Another aspect leading to ambiguous accuracy istha t the results are based onNDV I and other environmental variables, such as (e.g. average maximum temperature, rainfall, altitude) where the independent marginal effect of each variable isunknown . However, this conditional effect can be as high as 50t o 75 % of the explained variation (Baylis etal, 1998;Robinso n et al, 1997). The major drawback ofusin g vegetation NDVI, however, isth e asymptotical approach to a saturation level above a certain biomass density and leaf area index (Tucker, 1977; Sellers, 1985;Tod d etal, 1998;Ga o etal, 2000;), and hastherefor e limited value in assessing biomass during, for example, thepea k of seasons (Thenkabail et al, 2000). This problem could be overcome by using more recent remote sensing products and techniques, such as the enhanced vegetation index (EVI) from the MODISproduc t suite (Huete et al, 2002). The EVI was developed to optimize the vegetation signal with improved sensitivity in high biomass regions and improved vegetation monitoring through a de-coupling of the canopy background signal and areductio n in atmosphere influences. Further, several studies explored the possibilities of narrow band vegetation indices for biomass estimation at high canopy density: Todd etal. (1998 ) and Clevers and Jongschaap (2001)reporte d a widening and deepening ofth e red absorption pit with an increase in biomass. Mutanga and Skidmore (2004a) successfully estimated biomass based on band depth analysis for densely vegetated areas where NDVI values reached an asymptote. In a similar approach itwa s demonstrated that imaging spectrometer data enhance the estimation of forest stand variables (leaf area index and crown volume) compared to broadband multispectral data. (Schlerf etal, 2005). 142 As indicated above, novel indices developed asproxie s for biomass and primary productivity, which arebased on spectrometer data, encompass the capability of moving beyond conventional NDVI analysis,particularl y suitable for identifying habitats in heterogeneous, densely vegetated areas. Assessing speciesrichnes s through temporal heterogeneity. Ecoclimatic dynamics are highly complex, ranging from the impact of changing weather, seasonal variation in climate, including interannual cycles,t o climate changes such as the global Pleistocene glacial periods. Seasonal variations in climate govern differences in plant species growth and establishment patterns, leading to changes in species composition and distributions (Hobbs, 1990). Consequently, annual variations in vegetation can induce changes in the spatial distribution ofplan t phenology and growth (Tucker and Selles, 1986). Therefore, analysis of multi-annual land cover data potentially provides a key to understanding the influence of climate variability on shaping ecosystems -whic h form the overarching hierarchical layer inbiodiversit y assessment. Continuous data to study ecoclimatic dynamics are available from 1980,wit h the establishment ofth e AVHRR meteorological satellite series. The coefficient ofvariatio n of AVHRR-derived NDVI data for anumbe r ofyear s indicates the relative variability of the vegetation cover for a given region. Consequently, regions with ahig h coefficient of variation should reflect large variations invegetatio n composition and growth, following unstable and unpredictable climatic conditions over anumbe r ofyears . Onth e other hand, low coefficients of variation should indicate regions with small variations in vegetation composition and growth. The use of temporal heterogeneity as aprox y is limited to a small number of studies, therefore all taxa are treated within the same chapter. Species richness and abundance of large mammals in Kenya were correlated with yearly variation in vegetation, as assessed by the interannual variation ofth e maximum AVHRR- derived NDVI (Oindo 2002). In linewit h the findings above,maximu m numbers of species were found in regions with current ecoclimatic stability. These studies support the 143 hypothesis that high species diversity occurs in stable,predictabl e environments (Sanders and Hessler, 1969;Fjelds a and Lovett, 1997).Fjelds a etal. (1997 ) correlated interannual variability ofNDV I with biodiversity 'hotspots' of tropical Africa, linking local endemism with local ecoclimatic stability. Similarly, Fjeldsa etal. (1999) correlated interannual differences inNDV I with endemism of 789Andea n bird species, thereby linking biodiversity with short-term and long-term ecoclimatic stability. Their results suggest that high ecoclimatic stability allows species to 'accumulate' in an area, whereas large interannual variation limitsth e community to species able to withstand large fluctuations in habitat quality resulting from interannual climatic variation. Rodriguez etal. (2005 ) used regression analyses to examine the relationship between reptile and amphibian species richness and a set of environmental variables related to five hypotheses for geographical patterns of species richness based onproductivity , ambient energy, water-energy balance, habitat heterogeneity, and climatic variability. For reptiles, annual potential evapotranspiration (an index of atmospheric energy) explained 71% of the variance. For amphibians, annual actual evapotranspiration (an index of thejoin t availability of energy and water inth e environment), and the global vegetation index derived from satellite data, both described about 60%o f the variance. Their results were consistent with reptile and amphibian environmental requirements, where the former depend strongly on solar energy, and the latter on both warmth and moisture for reproduction. On a somewhat different vein, Carey etal. (2001 ) attempted to identify the causes of amphibian declines around the globe. Four relatively undisturbed areas in northeastern Australia, Costa Rica-Panama, central Colorado, and Puerto Rico were chosen for examination of environmental correlates coincident with mass mortalities at these localities. They compiled a database including descriptions of 120 localities,bot h at which declines have been documented and atwhic h no declines were known atth e time. For each locality, the number of species, dates and degree of declines, habitat characteristics, and other factors were provided. The authors then used data predicted by models or collected by satellites, airplanes, or direct sampling on the ground to evaluate variations over time in temperature, precipitation, wind direction, UV-B radiation, and concentrations of 144 contaminants. They considered the variation in certain environmental variables unlikely to have directly caused amphibian deaths,bu t suggested that correlations between these environmental changes and the occurrence of amphibian die-offs need further investigation into synergistic interactions among environmental variables andpossibl e indirect causal relationships. Summary This approach assesses species richness based on temporal heterogeneity, Themulti - temporality ofhabita t heterogeneity, indicated through the heterogeneity inNDVI , was suggested tob e an appropriate proxy to predict species richness patterns (Sanders and Hessler, 1969;Fjelds a and Lovett, 1997). The literature review showed that temporal studies were mostly performed across large regional areas.Further , conflicting results highlight the need to select relevant taxa (refer to taxonomic resolution) and to tune the methodology (Oindo and Skidmore, 2002). Even though NOAA-AVHRR data currently offer the longest time series,the y are limited inthei r spatial and spectral resolution. The variability in vegetation cover as assessed using AVHRR data isth e result of multiple influences: intrinsic characteristics of climate such as interannual variability in rainfall and temperature, long-term climate trends, vegetation succession, anthropogenic land-cover changes, and variability inth e state of atmosphere (Fjeldsa etah, 1997).However , the prospective for the temporal analysis ispromisin g as alternative long-term satellite data series evolve. In the future, the combination ofmultitempora l satellite data with historical meteorological, ecological andpaleologica l data has the potential of describing interactions among seasonal, annual and long-term climate variability to understand species diversity. Multi-temporal data offer possibilities to overcome the limitations of 'static' habitat studies needed for conservation purposes. Given the fact that many species are extremely mobile over time, e.g. migratory species, single-date studies do not cover the complete range of their habitats. In such cases only multitemporal data canprovid e amor e complete assessment of the species' occurrence and distribution. 145 Assessing species richness through heterogeneity based on landscape structural properties Brokaw and Lent (1999) stated that, in general, the more vertically diverse a forest is,th e more diverse is itsbiota . Remote sensing has the potential to estimate structural properties and assess their heterogeneity. Most studies relating remote sensing derived structural properties to animal diversity relied on height measuring technologies such as airborne lasers (i.e., airborne LiDAR) and Synthetic Aperture Radar (SAR). They are tools to map vegetation height and itsvariability , field boundary height, shape of individual agricultural fields, fractional vegetation cover, and aboveground biomass (e.g., Ritchie etal, 1995; Blair etal, 1999;Lefsk y etal, 2002; Lim etal, 2003; Mason et al, 2003, Santos et al, 2003;Lefskyefa/.,2005). Recently,Nelso n etal. (2005 ) analyzed LiDAR measurements and videot o identify and locate forested sites that might potentially support populations of amammal , the Delmarva fox squirrel (Sciurus niger cinereus). Results indicated that the largest part of the area (78%) met certain minimum length, height, and canopy closure criteria to support squirrel populations. This isth e only study addressing a faunal taxa other than the avifauna. Beier and Drennan (1997) demonstrated thatNorther n Goshawks (Accipiter gentilis) selected foraging sites based on structure rather than onpre y abundance, while (Jansson and Andren, 2003) found that forest structure isrelate d to species richness. Imhoff et al, (1997)use d SAR and aerial photography to map vegetation heterogeneity and relate this to field studies of bird abundances in Australia's Northern Territory. The abundances of individual species changed significantly across floristic and structural gradients, implying that bird habitat can bepredicte d from SAR data. Hinsley etal. (2002 ) and Hill et al. (2003) used an airborne laser scanning (ALS) system to map forest structure and related canopy heights to chick mass (i.e.,nestlin g weight), a surrogate for breeding success, which, inturn , isa function of 'territory quality'. They found that, for one species, chick mass increased with increasing forest canopy height, and for a second species, chick mass 146 decreased. Hence,Hill etal. (2003 ) concluded that airborne laser scanning data canb e used topredic t habitat quality and to map species distributions as a function ofhabita t structure. Davenport etal. (2000) devised atechniqu e to measure the height of crops in farmland fields using LiDAR scanning, as crop height is an important predictor of bird species population and, inturn , can be used as aprox y for bird suitability. Using a population model of skylark {Alaudaarvensis) they concluded thatth e achieved structural accuracy - less than 10c m -woul d be sufficient to discriminate suitable from unsuitable habitat from LiDAR data. Incorporating high resolution multi-spectral images,thes e techniques can be used over large geographical areas and couldtherefor e havewid e application in ecological monitoring of change inhabita t structures andth e associated effects onwildlif e (Mason et al, 2003). Summary The fourth approach, assessing species richness through heterogeneity based on landscape structural properties, involves the assessment of species diversity using structural properties of habitat heterogeneity. This isa more complicated approach that not only relies on primary productivity and itsheterogeneity , but also on structural properties of ecosystems. Vegetation height, height variability, percent canopy cover, and aboveground biomass are structural properties defining habitat heterogeneity (e.g. Ritchie etal, 1995;Blai r et al, 1999;Lefsk y etal, 2002,2005 ;Li m etal, 2003; Santos etal, 2003). Such structural habitat properties were successfully correlated to species distributions (Imhoff etal, 1997; Brokaw and Lent, 1999;Hinsle y etal, 2002;Jansso n and Andren, 2003; Hill etal, 2003). In addition to the applied laser and radar systems, optical multiangular sensor products (e.g., from the Multiangular Imaging SpectroRadiometer (MISR)) describing structural properties, such asth e leaf area index and leaf angle distribution haveth epotentia l to contribute to the assessment and monitoring ofterrestria l faunal species richness. In addition to more traditional approaches, the use of structural characteristics ofhabitats , their change and influence on faunal species distribution has ahig h potential for further studies covering large geographical areas (Nelson etal, 2005). 147 Assessing species richnessthroug h heterogeneity based onplan t chemical constituents Animal species have apreferenc e for the spatial and structural composition ofhabitat , but another attractant is the forage quality that an animal perceives in that habitat. Studies in the African savannah demonstrated that the occurrence and spatial distribution of many wildlife species is influenced by the variation in grass quality (Grant et al, 2002;Heitkoni g and Owen-Smith, 1998;McNaughton , 1988). Techniques that can estimate canopy quality on a large scale appear relevant in understanding wildlife diversity. Broadband satellites such as Landsat TM or SPOT lack the potential to capture detailed spectral features needed to detect or estimate the concentration of chemical constituents. Alternatively, imaging spectrometers can measure canopy reflectance in narrow and contiguous spectral bands in a wide wavelength range (e.g.40 0 - 2500 nm).A wide range ofplan t compounds and their concentration can be identified from the many subtle absorption features of the spectrometer data (Curran, 1989;Elvidg e 1990). Therelationship s with spectral properties and foliar chemicals, nitrogen amongst others,hav ebee n studied from dried and fresh leaves (e.g. Grossman etal, 1996;Dur y and Turner, 2001;Ebber s etal, 2002), to entire canopies (e.g.Jago etal, 1999;Curra n etal, 2001). However, there are many complicating factors to consider when estimating biochemicals of entire canopies. These include the masking effect of leaf water absorption (Fourty and Baret, 1998), the complexity of the canopy architecture, variation in leaf internal structure and directional, atmospheric and background effects. Several methods were developed to maximize sensitivity to the vegetation characteristics while minimizing confounding factors, including band ratios, difference indices, and derivative analysis (e.g., Huang etal 2004; Schmidt and Skidmore, 2004; Becker etal, 2005). Regarding forage quality assessment as proxies for animal studying, Mcllwee et al.(2001 ) investigated the utility of in situ reflectance spectroscopy as a means of rapidly assaying chemical constituents of leaves of four Eucalyptus species topredic t herbivory by greater gliders {Petauroidesvolans) and common ringtail possums (Pseudocheirus peregrinus). 148 Resulting concentrations of nitrogen, neutral detergent fibre, condensed tannins and total phenolics, and thus leaf palatability, were predicted accurately and were consistent with documented food preferences of greater gliders. Dury etal. (2001 ) estimated concentrations of nutrients in Eucalypt tree foliage using airborne imaging spectrometer data. They determined secondary compounds ofth e group diformylphloroglucinols known to be deterrents for herbivores. Consequently, they derived palatability of Eucalypt leaves for foliovorous marsupials to mappotentia l koala andpossu m habitats. Mutanga etal. (2004c ) investigated the ability of field spectroscopy to discriminate different levels of sodium concentration in grass, as sodium is a scarce element needed and sought by mammals (e.g. Brady etal., 2002; Grant etal, 2002).Usin g field spectrometer measurements ofpastur e grass, they were able to detect several sodium concentrations. They concluded that with the knowledge of grass species distribution, imaging spectrometer data would help to understand the distribution of mammals in nutrient limited savannas. This approach was successfully applied to large geographical areas linking forage quality to species richness and distribution particularly in areas with limited nutrients (Ferwerda, 2005). Summary This last approach isbase d on the use of plant chemical constituents to define habitat heterogeneity and ultimately assess and predict species richness. Attractants and deterrents related to the structural and trophic composition ofhabita t are important criteria to be considered inhabitat-specie s associations. These attractants canb e forage quality (Grant et al, 2002;Heitkoni g and Owen-Smith, 1998;McNaughton, 1988).Consequentl y the estimation of forage quality is essential to understand species richness patterns. Imaging spectroscopy with its ability to record reflected radiance in narrow spectral bands, allows the detection and quantification of canopy biochemical components. The overview of these initial studies demonstrated the utility of imaging spectrometer data to map foliar nutrient concentration in savannas and woody ecosystems. The above mentioned studies provide a 149 first step towards understanding the movement and distribution of wildlife, particularly in areas where herbivorous wildlife is known to be limited by nutrients. The correlations between animal presence-abundance/habitat and forage quality were consistent with results derived from imaging spectroscopy; Mcllwee etal., 2001; Mutanga etal, 2004c d),whils t other studies successfully upscaled those correlations to large geographical areas (Dury et al., 2001, Mutanga etal., 2004b) ,particularl y in areas with limited nutrients (Ferwerda, 2005). Some authors recommended that future studies should focus on monitoring seasonal changes in foliar nutrient concentration aswel l as extending the method topredic t other macro nutrients (P,K ,Na , Mg, Ca) and secondary compounds in both grass and tree canopies.Nevertheles s a major constraint remains that foliar chemicals contribute only a littlet o the canopy optical properties.Radiativ e transfer models incorporating the involved optical mechanisms at varying complexity have some success at biochemical parameter retrieval (e.g.Jacquemou d etal, 2000;Kot z etal. 2004) , however typically many inputs of canopyparameter s are required. At the current stage extensive in situ investigations on spectral features of attractants and deterrents of forage and their influence on faunal species distributions is aprerequisit e to successfully upscale these findings to large areas for monitoring and conserving faunal species. Conclusion It is important for conservation purposes to generate consistent and reliable information about species distribution and diversity in order to develop plans for species protection and sustainable use (Riede,2000) . Remote sensing is generally regarded to be able to contribute to this aim, mainly by its ability to provide continuous spatial information. It is rapidly developing, capable of coping with environmental heterogeneity -an d thus biodiversity - in itsbroades t range, by measuring increasingly detailed variation on a spatial,temporal , and structural scale, and recently by measuring variation in biochemical composition. This development isclearl y highlighted in thepresente d review. 150 A major issue complicating the assessment of species occurrence and richness across all techniques isth e mobility of faunal species, especially migrants which can move long distances occupying awid e range ofnatura l and anthropogenic habitats. Techniques used in plant oriented diversity studies are generally based on characteristic spectral reflectance features of plant species orplan t communities. For this purpose, objects need tob e sessile to be accurately assessed. Thetechnique s ofremot e sensing that have aided the studies on plant species distribution and diversity cannot be applied to animal studies in a similar fashion. Although commercial satellites with 61 cm pixel size are now capable of locating elephants in Kenya's Amboseli National Park and surrounding ecosystem (see e.g. http://media.digitalglobe.com/file.php/binaries/51/AmboseliF.pdf), most animal species remain undetectable. The often cryptic existence of fauna poses an additional general problem, akin to that ofundergrowt h species in vegetation. Table 1 presents a summary of thevariou s approaches discussed inthi s paper, including aspects of methodology, data requirements, techniques involved, and the biodiversity aspects covered. Most of the earlier studies on the application of remote sensing to biodiversity research were published in ecological, rather than remote sensingjournals . This review is mainly based on peer-reviewed literature. Two main biases can be identified in the selected studies: a strong emphasis onNDVI-base d approaches and on spaceborne sensor data. The development and application of new remote sensing techniques and products appears to undergo a lag time before they enter the realm of ecological research. We acknowledge thatne w methodologies (e.g.,narrow-ban d vegetation indices), and less operational systems (e.g., airborne imaging and in situ spectroradiometers, lidar)hav e a high potential to provide ane w generation ofvegetatio n products representing proxies to estimate animal distribution. 151 •N* > -Si •- E ^3 Q. o d. o If e -s; 5f 6 o I o U o O O H H O O •« < CJ o< 5b z a. U T3 ^~v o ^3 in oog> s s •Si op J o Finally, this review has focussed on the use ofremot e sensing for estimating terrestrial animal distribution and diversity. Geostatistical methods are more and more being used incorporating remote sensing and field measurements. Theus e of ancillary data such as climate,terrain , soils,huma n infrastructure and footprint, access to water and so on, have been extensively used in geographic information systems. Many of these GIS modelling exercises also incorporate remotely sensed imagery. Corsi et al (2000) provide an overview of these models and ancillary data sources. Acknowlegments The authors wish to acknowledge Michael Schaepman and Jan Clevers for including the review article inthi s special issue, andtw o anonymous reviewers for their comments. 153 References Asner, G.P.,Vitouse k P.M., 2005.Remot e analysis of biological invasion and biogeochemical change. PNAS, 102(12) ,4383-4386 . Arntzen, J.W., Alexandrino, J., 2004. Ecological modelling of genetically differentiated forms ofth e Iberian endemic golden-striped salamander, Chioglossa lusitanica. Herpetol.J. 14, 137. Austin, G.E., Thomas, C.J., Houston, D.C., Thompson, D.B.A., 1996.Predictin g the spatial distribution of buzzard Buteo buteo nesting areas using a geographical information system and remote sensing. J. Appl. Ecol. 33, 1541-1550. Avery, M.I., Haines-Young, R.H., 1990.Populatio n estimates for the dunlin Calidris alpina derived from remotely sensed satellite imagery ofth e Flow Country of northern Scotland. Nature 344, 860-862. Bailey, S.A., Horner-Devine, M.C., Luck, G., Moore, L.A., Carney, K.M., Anderson, S., Betrus, C, Fleishman, E., 2004. Primary productivity and species richness: relationships among functional guilds, residency groups and vagility classes at multiple spatial scales.Ecograph y 27 (2),207-217 . Baines, L. M., Collins, W. G., Robins, P., 1986. The application ofremot e sensing to urban bird ecology. Remote sensing for resources development and environmental management. Proc. 7th ISPRS Commission VII symposium, Enschede, Vol. 2(1) : 795-799. Baylis, M., Rawlings, P., 1998.Modellin g the distribution and abundance ofCulicoides imicola in Morocco and Iberia using climatic data and satellite imagery. Arch.Virol .14 , 137-153. Baldi, A., 2003. Using higher taxa as surrogates of species richness: a study based on 3700 Coleoptera, Diptera, and Acari species in Central-Hungarian reserves. Basic Appl. Ecol.4 ,589-593 . Barnsley, M.J., Hobson, P.,Lewis , P., O'Dwyer, S., Disney, M.I., Cutter, M., Lobb,D. , 2000. On the potential of CHRIS/PROBA for estimating vegetation canopy properties from space. Rem. Sens.Rev.1 9 (1-4), 171-189. 154 Bavia, M.E., Malone,J.B. ,Hale , L.,Dantas ,A. , Marroni, L., Reis, R., 2001. Use of thermal and vegetation index data from earth observing satellites to evaluate theris k of schistosomiasis in Bahia, Brazil. Acta Trap. 79 (1),79-85 . Bechtel, R., Sanchez-Azofeifa, A., Rivard, B.,Hamilton , G., Martin, J., Dzus, E.,2004 . Associations between Woodland Caribou telemetry data and Landsat TM spectral reflectance. Int. J. Rem. Sen. 25 (21), 4813-4821. Beier, P.,Drennan , J.E., 1997.Fores t structure and prey abundance in foraging areas of Northern Goshawks. Ecol. Applic. 7 (2), 564-571. Blair J.B.,Rabine ,D.L. , Hofton, M.A., 1999.Th e Laser Vegetation Imaging Sensor: A medium-altitude, digitisation-only, airborne laser altimeter for mapping vegetation and topography. ISPRS J. Photogram. Rem. Sen. 54 (2-3), 115. Box, E.O., Holben, B.N. ,Kalb , V., 1989.Accurac y of the AVHRR Vegetation Index as a predictor ofbiomass , primary productivity and net C02 flux. Vegetatio 80,71 - 89. Brady, R.C., Track, S.C., Jones, C.J., Ross,P.F. ,2002 . Stunted growth and mortality associated with sodium deficiency. Avian Dis.46 , 740-744. Brokaw N.V., Lent, R.A., 1999.Vertica l structure, in Hunter, M. (Editor). Maintaining biodiversity in forest ecosystems. Cambridge University Press, Cambridge, pp. 373-399. Brooks, T., Balmford, A., Burgess,N. , Fjeldsa, J., Hansen, L.A., Moore,J. , Rahbek, C, Williams, P., 2001. Toward ablueprin t for conservation in Africa. BioScience 51(8), 613-624. Brown, J.H., 1988.Specie s diversity, in Myers, A.A . and P. S., Giller, (Editors). Analytical Biogeography: An Integrated Approaches toth e Study of Plant and Animal Distribution Chapman &Hall ,Ne w York,pp . 57-89. Brown, J.H., Lomolino, M.V., 1998.Biogeography , 2nd ed. Sinauer Associates, Sunderland, Massachusetts, 691p . Cardillo,M. , Macdonald, D.W., Rushton, S.P., 1999.Predictin g mammal species richness and distributions:testin g the effectiveness of satellite-derived land cover data. Landscape Ecol. 14(5) ,423-435 . 155 Carey, C, Heyer, W.R., Wilkinson, J., Alford, R.A., Arntzen, J.W., Halliday, T., Hungerford, L., Lips, K.R., Middleton, E.M., Orchard, S.A., Rand, A.S., 2001. Amphibian declines and environmental change: Use of remote-sensing data to identify environmental correlates. Conserv. Biol. 15(4) , 903-913. Catling P.C., Burt, R.J., Forrester, R.I., 2000. Models ofth e distribution and abundance of ground-dwelling mammals in the eucalypt forests of north-eastern New South Wales in relation to habitat variables. Wildlife Res. 27 (6),639-654 . Ceballos, G., Ehrlich, P.R., Soberon, J., Salazar, I., Fay,J.P. ,2005 .Globa l mammal conservation: What must we manage? Science 309,603-607 . Chust, G., Pretus,J.L.I. , Ducrot, D., Ventura, D.,2004 . Scale dependency of insect assemblages inrespons e to landscape pattern. Landscape Ecol. 19(1) ,41-57 . Clevers, J.G.P.W., Jongschaap, R., 2001. Imaging spectrometry for agricultural applications, inVa n der Meer, F. D.,an d S. M., De Jong, (Editors). Imaging spectrometry: Basic principles and prospective applications. Kluwer Academic Publishing. Dordrecht, pp. 175-179. Coops,N.C. ,Catling , P.C., 1997.Predictin g the complexity of habitat in forests from airborne videography for wildlife management. Int. J. Rem. Sen. 18(12),2677 - 2682. Coops,N. , Culvenor, D.,Preston , R., Catling, P., 1998.Procedure s for predicting habitat and structural attributes in eucalypt forests using high spatial resolution remotely sensed imagery. Aust. Forestry 61 (4),244-252 . Coops,N.C. ,Catling , P.C., 2002.Predictio n of the spatial distribution and relative abundance of ground-dwelling mammals using remote sensing imagery and simulation models. Landscape Ecol. 17(2) , 173-188. Corsi, F.,d e Leeuw, J., Skidmore,A.K. , 2000. Modelling species distribution with GIS. In: Boitani, L., Fuller, T.K. (Eds.), Research Techniques in Animal Ecology Controversies and Consequences. Columbia University Press,Ne w York. pp.389 - 434. 156 Cowley, M.J.R., Wilson, R.J., Leon-Cortes, J.L., Gutierrez, D., Bulman, C.R., Thomas, CD., 2000. Habitat-based statistical models for predicting the spatial distribution of butterflies and day-flying moths in a fragmented landscape. J. Appl. Ecol. 37 (1), 60-72. Cracknell, A.P., 1998.Synerg y in remote sensing -what' s in apixel ? Int. J. Rem. Sen.19 (11), 2025-2047. Curran, P.J., 1989. Remote sensing of foliar chemistry -Rem . Sen. Environ.. Rem. Sen. Environ. 30,271-278. Currie,D.J. , 1991. Energy and large-scale patterns of animal-and plant-species richness. Am.Nat . 137(11),27-49 . Cushman, S.A., McGarigal, K.,2004 .Pattern s in the species-environment relationship depend on both scales and choice of response variables. Oikos 105,117-124. Davenport, I.J., Bradbury, R.B.,Anderson , G.Q.A., Hayman, G.R.F., Krebs,J.R. , Mason, D.C.,Wilson , J.D., Veck,N.J. ,2000 . Improving bird population models using airborne remote sensing. Int. J. Rem. Sen. 13-14,2705-2717 . Debinski, D.M., Kindscher, K.,Jakubauskas , M.E., 1999.A remote sensing and GIS-based model of habitats and biodiversity inth e Greater Yellowstone Ecosystem. Int. J. Rem. Sen. 20 (17), 3281-3291. Del Valle, H.F., De Lamo, D.A., Gagliardini, D.A., 1997.Environmenta l affinity of the guanaco (Lama guanicoe M3ller, Camelidae) intw o selected areas of Central Patagonia supported by ERS-1 SAR data. Earth Obs. Quart. 55, 14-19. DeWulf , RR, Goossens, R.E., McKinnon, J.R. and Cai,W.S. , 1988.Remot e sensing for wildlife management: Giant panda habitat mapping from LANDSAT MSS images. Geocarto Int. 1,41-50. Diner D.J., Braswell, B.H. , Davies, R., Gobron, N.,Hu , J., Jin, Y.Kahn , R., Knyazikhin, Y., Loeb,N. , Muller, J.P.,Nolin , A., Pinty, B.,Schaaf , C. B., Seiz, G., Stroeve, J., 2005. The value of multiangle measurements for retrieving structurally and radiatively consistent properties of clouds, aerosols, and surfaces. Rem. Sen. Environ. 97,495-518. 157 Doerries, M.B.,Va nDover , C.L.,2003 . Higher-taxon richness asa surrogat e for species richness inchemosyntheti c communities. Deep-Sea Res.I 50 , 749-755. Dury, S.J., Turner, B.,Foley , W.J., 2001. Theus eo f high spectral resolution remote sensing todetermin e leaf palatability ofeucalyp t trees for folivorous marsupials. Int.J .Appl . Earth Observ.Geoinf. 3(4) ,328-336 . Eby, J.R.,Bright , L.R., 1985.A digita l GIS based onLandsa t andothe r data forel khabita t effectiveness analyis. inProceeding s ofth e 19thInternationa l Symposiumo n Remote Sensing ofth e Environment, Environmental Research Instituteo f Michigan, AnnArbor , ML,pp . 855-864. Eyre, M.D., Foster, G.N.,Luff , M.L., Staley,J.R. ,2003 . Aninvestigatio n intoth e relationship between water beetle (Coleoptera) distribution andlan d coveri n Scotland andnortheas t England. J.Biogeogr . 30(12) , 1835-1849. Eyre, M.D.,Luff , M.L., Staley,J.R. ,Telfer , M.G. 2003. Therelationshi p between British ground beetles (Coleoptera, Carabidae) andlan d cover. J.Biogeogr . 30(5) , 719- 730. Ferwerda, J.G., 2005.Chartin g thequalit y ofForage : measuring andmappin g thevariatio n of chemical components infoliag e with hyperspectral remote sensing.Ph D thesis, ITC, TheNetherlands . Fisher, P., 1997.The pixel:A snare anda delusion . Int.J . Rem. Sen. 18(3) ,679-685 . Fjeldsa, J.,Lovett , J.C., 1997.Biodiversit y andenvironmenta l stability. Biodivers.Conserv. 6,315-323. Fjeldsa, J.,Ehrlich , D., Lambin,E. ,Prins ,E. , 1997.Ar ebiodiversit y 'hotspots' correlated with current ecoclimatic stability? Apilo t study using theNEA-AVHR R remote sensing data. Biodivers. Conserv. 6(3) ,401-422 . Fjeldsa, J.,Lambin , E.,Mertens , B., 1999. Correlation between endemism andloca l ecoclimatic stability documented bycomparin g Andean bird distributionsan d remotely sensed land surface data. Ecography 22(1) ,63-78 . Fourty, T.,Baret , F. 1998.O nspectra l estimates offres h leaf biochemistry, Int.J . Rem. Sen. 19(7), 1283-1297. Franklin, W.L., 1983.Contrastin g socioecologies ofSout h America's wild camelids:Th e 158 vicuna and guanaco. Advances inth e Study of Mammalian Behaviour, in Eisenberg, J.F., and D.G., Kleiman, (Editors). The American Society of Mammalogists. Shippensburg, PA.pp . 573-629. Franklin, S.E., Raske, A.G., 1994. Satellite remote sensing of spruce budworm forest defoliation in western newfoundland. Can. J. Rem. Sen.20 (1),37-48 . Franklin, S.E., Hansen, M.J., Stenhouse, G.B., 2002. Quantifying landscape structure with vegetation inventory maps and remote sensing. Forest. Chron. 78 (6),866-875 . Fuller, R.M., Groom, G.B., Mugisha, S., Katende, A., Ipulet, P.,Pomeroy , D.,Bailey , R., Ogutu-Ohwayo, R., 1998. The integration of field survey and remote sensing for biodiversity assessment: A case study in the tropical forests and wetlands of Sango Bay, Uganda. Biol. Conserv. 86 (3), 379-391. Fuller, R.M., Devereux, B.J., Gillings, S., Amable, G.S.,Hill , R.A., 2005.Indice s of bird- habitat preference from field surveys ofbird s and remote sensing of land cover: A study of south-eastern England with wider implications for conservation and biodiversity assessment. Global Ecol. Biogeogr. 14(3) ,223-239 . Gao, X., Huete, A.R., Ni, W., Miura, T., 2000. Optical-biophysical relationships of vegetation spectra without background contamination. Rem. Sen. Environ. 74(3) , 609-620. Gaston, K.J., Blackburn, T.M., 2000. Patterns and process in macroecology. Blackwell Science, Oxford. Gould, W., 2000. Remote sensing ofvegetation , plant species richness, and regional biodiversity hotspots. Ecol. Appl. 10(6) , 1861-1870. Grant, C.C., Davidson, T., Funston, P.J., Pienaar, D.J., 2002. Challenges faced in the conservation of rare antelope: A case study on the northern basalt plains of the KrugerNationa l Park. Koedoe 45 (2), 1-22. Green, K.M., Lynch, J.F., Sircar, J., Greenberg, L.S.Z., 1987.Landsa t remote sensing to assess habitat for migratory birds in the Yucatan Peninsula, Mexico.Vid a Silvestre Neo-Trop. 1,27-37. 159 Griffiths, G.H., Lee, J., Eversham, B.C., 2000. Landscape pattern and species richness; regional scale analysis from remote sensing. Int. J. Rem. Sen. 21 (13-14),2685 - 2704. Hall, R.J., Hall, R.J., Crown, P.H., Titus, S.J., Volney, W.J.A., 1995.Evaluatio n of LANDSAT thematic mapper data for mapping top kill caused by Jack Pine budworm defoliation. Can. J. Rem. Sen. 21 (4),388-399 . Hall, R.J., Hogg, E.H., Brandt, J.P., Case, B.S.,Fernandes , R.A., Butson, C, Leblanc, S.G., 2003. Relating aspen defoliation to changes in leaf area derived from field and satellite remote sensing data. Can. J. Rem. Sen. 29 (3), 299-313. Hawkins, B.A., Porter, E.E., 2002.Doe s Herbivore Diversity Depend on Plant Diversity? The Case of California Butterflies. Am.Nat . 161(1), 40-49. Hawkins, BA., 2004. Summer vegetation, deglaciation and the anomalous bird diversity gradient in eastern North America. Global Ecol. Biogeogr. 13(4) , 321-25. Hawkins, B.A., Porter, E.E., Diniz-Filho, J.A.F., 2003. Productivity and history as predictors ofth e latitudinal diversity gradient ofterrestria l birds. Ecology 84, 1608-1623. Hay, S.I., 1997.Remote sensing and disease control:Past , present and future. T. Roy. Soc. Trap. Med. H. 91 (2), 105-106. Hay, S.I., Rogers, D.J., Snow, R.W., 1998.Predictin g malaria seasons in Kenya using multitemporal meteorological satellite sensor data. T. Roy. Soc.Trop . Med. H. 92 (1), 12-20. Heino,J. , Soininen, J., Lappalainen, J., Virtanen, R., 2005.Th e relationship between species richness and taxonomic distinctness in freshwater organisms. Limnol. Oceanogr. 50 (3),978-986 . Heitkonig, I.M.A., Owen-Smith, N., 1998.Seasona l selection of soil types and grass swards by roan antelope in a South African savanna. African J. Ecol. 36 (1),57-70 . Heitkonig, I.M.A., Ringrose, S., Epema, G.F., Bonyongo, C, Veenendaal, E., Stein, A., Breeuwer, A., Hasselt, M. van, Klop, E., Goethem, J., Slot, M., 2003. Watching Wildlife from Space. In: Bernard, T., Mosepele, K., Ramberg, L. (Eds.), Int. Wetlands Conference, Botswana. 160 Hepinstall, J.A., Sader, S.A., 1997.Usin g Bayesian statistics,themati c mapper satellite imagery, and breeding bird survey data to model bird species probability of occurrence in Maine. Photogramm. Eng. Rem. Sen. 63 (10), 1231-1237. Herr, A.M., Queen, L.P., 1993.Cran e habitat evaluation using GIS and remote sensing. Photogramm. Eng. Rem. Sen. 59, 1531-1538. Hill, R.A., Hinsley, S.A., Bellamy, P.E., Balzter, H., 2003. Ecological applications of airborne laser scanner data: Woodland bird habitat modeling, in Hyypa, J.,Nassset , E., Olsson, H., Granqvist Pahlen, T. andH. , Reese, (Editors) Proceedings of the Scandlaser Scientific Workshop on Airborne Laser Scanning of Forests,Umea , Swedish University of Agricultural Sciences,pp . 78-87. Hinsley, S.A., Hill, R.A., Gaveau, D.L.A., Bellamy, P.E.,2002 . Quantifying woodland structure and habitat quality for birds using airborne laser scanning. Funct. Ecol. 16(6), 851-857. Hobbs, R.J., 1990,Remot e sensing of spatial and temporal dynamics of vegetation, in Hobbs, R. J., and H.A. , Mooney, (Editors). Ecological Studies 79,Rem . Sen. of Biosphere Funct., New York: Springer-Verlag, pp.203-219 . Homer, C.G., Edwards, T.C. Ramsey, R.D., Price, K.P., 1993.Us e ofremot e sensing menthods in modelling sage grouse winter habitat. J. Wildlife Manage. 57,78-84 . Huang, Z., Turner, B.J., Dury, S.J., Wallis, I.R., Foley, W.J., 2004. Estimating foliage nitrogen concentration from HYMAP data using continuum removal analysis. Rem. Sen. Environ. 93 (1-2), 18-29. Huber, T.P., Casler, K.E., 1990.Initia l analysis of Landsat TM data for elk habitat mapping. Int. J. Rem. Sen. 11(5) ,907-912 . Huete,A. , Didan, K., Miura, T., Rodriguez, E.P., Gao,X . and Ferreira, L. G., 2002. Overview ofth e radiometric and biophysical performance of the MODIS vegetation indices. Rem. Sen. Environ. 83 (1-2), 195-213. Hurlbert, H., Haskell, J.P., 2003. The Effect of Energy and Seasonality on Avian Species Richness and Community Composition. Am.Nat . 161, 83-97. Huston, 1994.Biologica l Diversity: The Coexistence of Species on Changing Landscapes. Cambridge University Press, Cambridge, 701 p. 161 Imhoff, M.L., Sisk, T.D., Milne, A., Morgan, G., Orr, T., 1997.Remotel y sensed indicators of habitat heterogeneity: Use of synthetic aperture radar inmappin g vegetation structure and bird habitat. Rem. Sen. Environ. 60 (3),217-227 . Ito, T.Y., Tsunekawa, A., Miura, N., Lhagvasuren, B.,Enkhbileg , D., Takatsuki, S., Jiang, Z., 2005.Preliminar y evidence of abarrie r effect of a railroad on the migration of Mongolian gazelles. Conserv. Biol. 19(3) ,945-948 . Jansson, G., Andren, H., 2003. Habitat composition and bird diversity in managed boreal forests. Scand. J. Forest Res. 18(3) ,225-236 . Johnson, C.G., Clausnitzer, R.R., Mehringer, P.J., Oliver, CD., 1994. Biotic and abiotic processes of eastside ecosystems: the effects of management on plant and community ecology, and on stand and landscape vegetation dynamics. U.S. Forest Service General Technical Report PNW-GTR-322, Pacific Northwest Research Station, Portland, Oregon. Johnson, D.D.P., Hay, S.I., Rogers, D.J., 1998.Contemporar y environmental correlates of endemic bird areas derived from meteorological satellite sensors. P. Roy. Soc. Lon. B.Bio . 265 (1400),951-959 . J0rgensen, A.F.,Nohr ,H. , 1996.Th e use of satellite images for mapping of landscape and biological diversity in the Sahel. Int. J. Rem. Sen. 17(1) ,91-109 . Joria, P.E., Ahearn, S.C., Connor, M., 1991. A comparison of the SPOT and Landsat Thematic Mapper satellite systems for detecting gypsy moth defoliation in Michigan. Photogramm. Eng. Rem. Sen. 57 (12), 1605-1612. Kanai, Y., Kondoh, A. Higuchi, H., 1994.Analysi s of crane habitat using satellite images. In: Higuchi, H., Minton, J. (Eds.) The Future of Cranes and Wetlands, Tokyo. Japan, pp. 72-81. Kokaly, R.F.,Despain , D.G., Clark, R.N., Livo,K.E. , 2003. Mapping vegetation in Yellowstone National Park using spectral feature analysis of AVIRIS data. Rem. Sen. Environ. 84 (3),437-456 . KQtz,B. , Schaepman, M., Morsdorf, F., Bowyer, P.,Itten , K., and Allgower, B.,2004 . Radiative Transfer Modeling within a Heterogeneous Canopy for Estimation of Forest Fire Fuel Properties. Rem. Sen. Environ. 92(3): 332-344. 162 Kristensen, T.K., Malone, J.B.,McCarroll , J.C., 2001. Use of satellite remote sensing and geographic information systems to model the distribution and abundance of snail intermediate hosts in Africa: Apreliminar y model for Biomphalaria pfeifferi in Ethiopia. Acta Trop. 79 (1),73-78 . Kumar, L., Schmidt, K.S.,Dury , S., Skidmore, A.K., 2001. Imaging spectrometry and vegetation science, inVa n der Meer, F.,an d S.M., De Jong, (Editors). Imaging spectrometry. Kluwer Academic Publishers, Dordrecht, pp. 111-115. Lack, D., 1969.Th enumber s of bird species on islands. Bird Study 16, 193-209. Laurent, E.J., Shi,H. , Gatziolis, D.,LeBouton , J.P., Walter, M.B.,Liu , J., 2005.Usin g the spatial and spectralprecisio n of satellite imagery topredic t wildlife occurrence patterns. Rem. Sen. Environ. 97,249-262 . Lauver, C.L., 1997. Mapping species diversity patterns in the Kansas shortgrass region by integrating remote sensing and vegetation analysis. J. Veg. Sci. 8(3) , 387-394. Lawrence, R., Labus, M., 2003. Early detection ofDouglas-fi r beetle infestation with subcanopy resolution hyperspectral imagery. West. J. Appl. For. 18(3) ,202-206 . Lee, P.F.,Ding , T.S., Hsu, F.H., Geng, S., 2004. Breeding bird species richness in Taiwan: Distribution on gradients of elevation, primary productivity and urbanization. J. Biogeogr. 31 (2),307-314 . Lefsky, M.A., Cohen, W.B.,Parker , G.G., Harding, D.J., 2002. Lidar Remote Sensing for Ecosystem Studies. BioScience 52 (1), 19-30. Lefsky, M.A., Hudak, A.T., Cohen, W.B.,Acker , S.A..,2005 .Geographi c variability in lidar predictions of forest stand structure inth e Pacific Northwest. Rem. Sen. Environ. 95 (4),532-548 . Lim, K., Treitz, P., Wulder, M., St-Onge, B.,Flood , M., 2003. LiDAR remote sensing of forest structure. Prog. Phys. Geog.2 7 (1),88-106 . Liu, J., Linderman, M., Ouyang, Z.,An , L.,Yang , J., Zhang, H., 2001, Ecological degradation inprotecte d areas: The case of Wolong Nature Reserve for giant pandas. Science 292, 98-101. Loffler, E., Margules, C, 1980.Wombat s detected from space.Rem . Sen. Environ. 9(1) , 47-56. 163 Loreau, M., 1998. Separating sampling and others effects in biodiversity experiments. Oikos 82 (3),600-602 . Luoto, M., Kuussaari, M., Toivonen, T., 2002.Modellin g butterfly distribution based on remote sensing data. J. Biogeogr. 29 (8), 1027-1037. Luoto, M., Virkkala, R., Heikkinen, R.K., Rainio, K., 2004. Predicting bird species richness using remote sensing in boreal agricultural-forest mosaics. Ecol. Appl. 14(6) , 1946-1962. Luther, J.E., Hudak, J., Meades, J.P., Franklin, S.E., 1997.Forecastin g the susceptibility and vulnerability ofbalsa m fir stands to insect defoliation with Landsat thematic mapper data. Rem. Sen. Environ. 59 (1), 77-91. Ma, J., Han, X., Hasibagan, A., Wang, C, Zhang, Y., Tang, J., Xie, Z.,Deveson , T.,2005 . Monitoring East Asian migratory locust plagues using remote sensing data and field investigations. Int. J. Rem. Sen. 26 (3),629-634 . MacArthur, R.H., Wilson, E.O., 1967.Th e theory of island biogeography. Princeton University Press, Princeton, 203p . Mace, G.M., 2005.A n index of intactness.Natur e 434: 32-33. Malone, J.B.,Yilma , J.M., McCarroll, J.C., Erko, B.,Mukaratirwa , S.,Zhou , X.,2001 . Satellite climatology and the environmental risk of Schistosoma mansoni in Ethiopia and east Africa. Acta Trop. 79 (1):59-72 . Mapaure, I.N., 1994. The distribution of Colophospermummopane (leguminosae - caesalpinioideae) in Africa. Kirkia 15, 1-5. Mason, D.C.,Anderson , G.Q.A., Bradbury, R.B.,Cobby , D.M., Davenport, I.J., Vandepoll, M., Wilson, J.D., 2003. Measurement of habitat predictor variables for organism- habitat models using remote sensing and image segmentation. Int. J. Rem. Sen. 24 (12),2515-2532 . McCoy, E.D.,Bell , S.S., 1991. Habitat structure: the evolution and diversification of a complex topic, in Bell, S.S., McCoy, E.D., and H.R., Mushinsky, (Editors). Habitat structure: the physical arrangement of objects in space. Chapman and Hall, London, pp. 3-27. 164 McCullough, D.G., 2000.A review of factors affecting the population dynamics ofjac k pine budworm {Choristoneurapinuspinus Freeman). Popul. Ecol.4 2 (3),243 - 256. Mcllwee, A.M., Lawler, I.R., Cork, S.J., Foley, W.J., 2001. Coping with chemical complexity in mammal-plant interactions:Near-infrare d spectroscopy as a predictor of Eucalyptus foliar nutrients and ofth e feeding rates of folivorous marsupials. Oecologia 128(4) ,539-548 . McNaughton, S.J., 1988.Minera l nutrition and spatial concentrations of African ungulates. Nature 334 (6180), 343-345. Millennium Ecosystem Assessment, 2005.Ecosystem s and Human Well-being: Biodiversity Synthesis.Worl d Resources Institute, Washington, DC, 85p . Miller, K.V., Conroy, M.J., 1990. SPOT satellite imagery for mapping Kirtland's warbler wintering habitat in the Bahamas.Wildlif e Soc. Bull. 18,252-257 . Minton, J.S.,Halls , J.H., Higuchi, H., 2003. Integration of satellite telemetry data and land cover imagery: a study of migratory cranes innortheas t Asia. T. in GIS 7 (4),505 - 528. Morin, PJ., 2000. Biodiversity's ups and downs.Natur e 406 (6795),463-464 . Morisson, R.I.G., 1997.Th e use ofremot e sensing to evaluate shorebird habitats and populations on Prince Charles Island, Foxe Basin, Canada. Arctic 50,55-75 . Murthy, M.S.R., Giriraj, A., Dutt, C.B.S., 2003. Geoinformatics for biodiversity assessment. Biol. Lett. 40 (2), 75-100. Musiega, D.E. and Kazadi, S.-N.,2004 . Simulating the East Afircan wildebeest migration patterns using GIS and remote sensing. African J. Ecol.42 ,355-362 . Mutanga, O., Skidmore, A.K., 2004a. Narrow band vegetation indices overcome the saturation problem in biomass estimation. Int. J. Rem. Sen. 25 (19), 3999-4014. Mutanga, O., Skidmore,A.K. ,2004b . Integrating imaging spectroscopy and neural networks to map grass quality in the Kruger National Park, South Africa. Rem. Sen. Environ. 90(1), 104-115. 165 Mutanga, O., Skidmore, A.K., Prins,H.H.T. , 2004c. Discriminating sodium concentration in a mixed grass species environment of theKruge r National Park using field spectrometry. Int. J. Rem. Sen. 25 (20), 4191-4201. Mutanga, O., Skidmore, AK, Prins H.H.T., 2004d. Predicting in situ pasture quality in the Kruger National Park, South Africa, using continuum-removed absorption features. Rem. Sen. Environ. 89 (3):393-408. Nagendra, H., 2001. Using remote sensing to assess biodiversity. Int. J. Rem. Sen. 22(12) , 2377-2400. Nagendra, H., Gadgil, M., 1999.Biodiversit y assessment at multiple scales: Linking remotely sensed data with field information. P.Natl .Acad . Soc. 96,9154-9158 . Negia, H.R., Gadgil, M., 2002.Cross-taxo n surrogacy ofbiodiversit y in the Indian Garhwal Himalaya. Biol. Conserv. 105, 143-155. Nelson, R.F., 1983.Detectin g forest canopy change due to insect activity using landsat MSS. Photogramm.Eng .Rem . Sen.4 9 (9), 1303-1314. Nelson, R., Keller C, Ratnaswamy, M., 2005.Locatin g and estimating the extent of Delmarva fox squirrel habitat using an airborne LiDAR profiler. Rem. Sen. Environ. 96 (3-4),292-301 . Nohr, H.,Jorgensen ,A.F. , 1997.Mappin g ofbiologica l diversity in Sahel by means of satellite image analyses and ornithological surveys. Biodiv. Conserv. 6 (4),545 - 566. Noss, R.F., 1990.Indicator s for monitoring biodiversity: Ahierarchica l approach. Conserv. Biol.4 (4):355-364 . Oindo, B.O.,2002 .Pattern s ofherbivor e species richness in Kenya and current ecoclimatic stability. Biodiv.Conserv. 11(7) ,1205-1221 . Oindo, B.O., Skidmore,A.K. , 2002. Interannual variability ofNDV I and species richness in Kenya. Int. J. Rem. Sen. 23 (2),285-298 . Oindo, B.O., Skidmore, A.K., De Salvo, P.,2003 . Mapping habitat and biological diversity in the Maasai Mara ecosystem. Int. J. Rem. Sen.2 4 (5), 1053-1069. Olme, C.D.L., Davies, R.G., Burguess, M., Eigenbrod, F.,Pickup ,N. , Olson, V.A., Webster, A.J., Ding, T-S.,Rasmussen ,P.C. ,Ridgley , R.S., Stattersfield, A.J., 166 Bennet, P.M., Blackburn, T.M., Gaston, K.J., Owens, P.F.,2005 .Globa l hotspots of species richness are not congruent with endemism orthreat . Nature 436, 1016- 1019. Olsgrad, F.,Brattegard , T., Holthe, T., 2003. Polychaetes as surrogates for marine biodiversity: lower taxonomic resolution and indicator groups. Biodiv. Conserv. 12, 1033-1049. Oppelt,N. , and Mauser, W., 2004. Hyperspectral monitoring ofphysiologica l parameters of wheat during a vegetation period using AVIS data. Int. J. Rem. Sen. 25 (1), 145- 159. Ormsby, J.P. and Lunetta, R.S., 1987.Whitetai l deer food availability maps from Thematic Mapper data. Photogramm. Engin.Rem. Sen. 53, 1585-1589. Osborne, P.E., Alonso, J.C., Bryant, R.G., 2001. Modelling landscape-scale habitat use using GIS and remote sensing: a case study with great bustards. J. Appl. Ecol. 38 (2), 458-471. Palmeirim, J.M., 1988. Automatic mapping of avian species habitat using satellite imagery. Oikos 52(1): 59-68. Perras, S., Bonn, F.,Gwyn , Q. H. J., Babin, M., Couture, R., 1988.Inventor y ofth e habitat ofth e American woodcock {Scolopaxminor) by remote sensing. Photo Interpretation: Images Aeriennes et Spatiales 88, 1-19. Price, K.P.,Jakubauskas , M.E., 1998. Spectral retrogression and insect damage in lodgepole pine successional forests. Int. J. Rem. Sen. 19(8) , 1627-1632. Prins, E., Petersen, B.S.,Aunins ,A. , Priednieks, J., 2005.Usin g Landsat TM and field data to produce maps of predicted bird densities in Latvian farmland. Int. J. Rem. Sen. 26(9), 1881-1891. Raxworthy, C.J., Martinez-Meyer, E., Horning, N.,Nussbaum , R.A., Schneider, G.E., Ortega-Huerta, M.A., Peterson, A.T., 2003. Predicting distributions of known and unknown reptile species in Madagascar. Nature 426, 837-841. Richards, B.J., 1999.Relationship s between white-tailed deer distribution and habitat characteristics in Hamilton County, Illinois.NCAS I Technical Bull. 2 (781),440 - 441. 167 Riede, K.,2000 . Conservation and modern information technologies: the global register of migratory species (GROMS).J . Int. Wildlife Law Policy 3 (2), 152-164. Riley, J.R., 1989.Remot e sensing inentomology . Annual Review of Entomology 34,247 - 271. Ritchie, J.C., Humes, K.S.,Weltz , M.A., 1995.Lase r altimeter measurements at Walnut Gulch watershed, Arizona. J. Soil Water Conserv. 50, 440-442. Robinson, T., Rogers, D.,Williams , B., 1997.Mappin g tsetse habitat suitability in the common fly belt of Southern Africa using multivariate analysis of climate and remotely sensed vegetation data. Med. Vet. Entomol. 11(3) ,235-245 . Rodrigues,A.S.L. , Andelman, S.J., Bakarr, M.I., Boitani, L., Brooks, T.M., Cowling, R.M., Fishpool, L.D.C., daFonseca , G.A.B., Gaston, K.J., Hoffmann, M., Long,J.S. , Marquet, P.A., Pilgrim, J.D., Pressey, R.L., Schipper, J., Sechrest, W., Stuart, S.N., Underhill, L.G., Waller, R.W., Watts, M.E.J., Yan, X., 2004. Effectiveness of the global protected area network in representing species diversity. Nature 428, 640-643. Rodriguez, M.A., Belmontes, J.A., Hawkins,B.A. , 2005.Energy , water and large-scale patterns ofreptil e and amphibian species richness in Europe. Acta Oecologica 28 (1), 65-70. Sader, S.A., Powell, G.V.N., Rappole, J.H., 1991. Migratory bird habitat monitoring through remote sensing. Int. J. Rem. Sen. 12,363-372 . Sanders, H.L., Hessler, R.R., 1969.Ecolog y of the deep-sea benthos. Science 163(874) , 1419-1424. Santos, J.R., Freitas, C.C., Araujo, L.S.,Dutra , L.V., Mura, J.C., Gama, F.F., Soler,L.S. , Sant'Anna, S.J.S., 2003. Airborne P-band SAR applied to the aboveground biomass studies in the Brazilian tropical rainforest. Rem. Sen. Environ.87 (4),482 - 493. Sauberer, N., Zulka, K.P., Abensperg-Traun, M., Berg, H.M., Bieringer, G., Milasowszky, N., Moser, D.,Plutzar , C, Pollheimer, M., Storch, C, 2004. Surrogate taxa for biodiversity in agricultural landscapes of eastern Austria. Biol. Conserv. 117(2) , 181-190. 168 Saveraid, E.H., Debinski, D.M., Kindscher, K., Jakubauskas, M.E., 2001. A comparison of satellite data and landscape variables inpredictin g bird species occurrences in the Greater Yellowstone Ecosystem, USA. Landscape Ecol. 16, 71-83. Saxon, E.C., 1983.Mappin g the habitats ofrar e animals in the Tanami wildlife sanctuary (Central Australia):A n application of satellite imagery. Biol. Conserv.27 (3),243 - 257. Schlerf, M., Atzberger, C, Hill, J., 2005.Remot e sensing of forest biophysical variables using HyMap imaging spectrometer data. Rem. Sen. Environ. 95 (2), 177-194. Scholes, R.J. and R. Biggs,2005 .A biodiversity intactness index. Nature 434:45-49 . Schwaller, M.R., Olson, Jr. C, Ma,Z. ,Zhu , Z., Dahmer, P., 1989.A remote sensing analysis ofAdeli e penguin rookeries. Rem. Sen. Environ. 28,199-206. Scribner, K.T., Arntzen, J.W., Cruddace,N. , Oldham, R.S.,Burke , T., 2001. Environmental correlates oftoa d abundance and population genetic diversity. Biol. Conserv. 98, 201-210. Sellers, P.J., 1985.Canop y reflectance, photosynthesis and transpiration. Int. J. Rem. Sen. 6 (8), 1335-1372. Sellers, P.J., 1987.Canop y reflectance, photosynthesis, and transpiration. II. Therol e of biophysics inth e linearity ofthei r interdependence. Rem. Sen. Environ. 21 (2), 143-183. Seto, K.C., Fleishman, E.,Fay , J.P., Betrus, C.J., 2004. Linking spatial patterns of bird and butterfly species richness with Landsat TM derived NDVI. Int. J. Rem. Sen. 25 (20), 4309-4324. Sharma, B.D., Clevers, J., De Graaf, R., Chapagain, N.R., 2004. Mapping Equus kiang (Tibetan Wild Ass) Habitat in Surkhang, Upper Mustang,Nepal . Mt. Res.Dev . 24 (2), 149-156. Simpson, E.H., 1949.Measuremen t of diversity.Natur e 163, 688. Sirois, J., Ahern, F.J., 1988.A n investigation of SPOT HRV data for detecting recent mountain pine beetle mortality. Can. J. Rem. Sen. 14(2) , 104-108. Smallie, J.J., O'Conner, T.G., 2000. Elephant utilization of Colophospermummopane: possible benefits of hedging.Africa n J. Ecol. 38,352-359 . 169 Smith, M.L., Martin, M.E., Plourde, L., Ollinger, S.V., 2003. Analysis of hyperspectral data for estimation oftemperat e forest canopy nitrogen concentration: Comparison between an airborne (AVIRIS) and a spaceborne (Hyperion) sensor. IEEE Transactions on Geoscience and Remote Sensing 41 (6)(I), 1332-1337. Stohlgren, T., Coughenour, M., Chong, G., Binkley, D., Kalkhan, M., Schell, L., Buckley, D., Berry, J., 1997. Landscape analysis of plant diversity. Landscape Ecol. 12: 155-170. Stoms, D.M., Estes, J.E., 1993.A remote sensing research agenda for mapping and monitoring biodiversity. Int. J. Rem. Sen. 14(10) , 1839-1860. Styles, C.V., Skinner, J.D., 1997. Seasonal variations inth e quality of mopane leaves as a source ofbrows e for mammalian herbivores. African J. Ecol. 35,254-265. Taft, O.W., Haig, S.M., Kiilsgaard, C, 2003. Use of radar remote sensing (radarsat) to map winter wetland habitat for shorebirds in an agricultural landscape. Environ. Manage. 32 (2), 268-281. Thenkabail, P.S., Smith, R.B.,D ePauw , E., 2000.Hyperspectra l vegetation indices and their relationships with agricultural crop characteristics. Rem. Sen. Environ. 71 (2), 158-182. Tilman, D., Lehman C.L., Thomson K.T., 1997.Plan t diversity and ecosystem productivity: Theoretical considerations. Proc.Natl .Acad . Sci.USA 94 (5), 1857- 1861. Tobler, M.W., Cochard, R., Edwards, P.J., 2003. The impact of cattle ranching on large- scale vegetation patterns in a coastal savannah in Tanzania. J. Appl. Ecol. 40(3) , 430-444. Todd, S.W., Hoffer, R.M., Milchunas, D.G., 1998.Biomas s estimation on grazed and ungrazed rangelands using spectral indices. Int. J. Rem. Sen. 19(3) ,427-438 . Tucker, C.J., 1977.Asymptoti c nature of grass canopy spectral reflectance. Appl. Optics 16 (5), 1151-1156. Tucker, C.J., 1979.Re d and photographic infrared linear combinations for monitoring vegetation. Rem. Sen. Environ. 8(2) , 127-150. 170 Tucker, C.J., Sellers. P.J., 1986. Satellite remote sensing ofprimar y production. Int. J. Rem. Sen. 7(11), 1395-1416. Verlinden, A., Masogo, R., 1997. Satellite remote sensing ofhabita t suitability for ungulates and ostrich in the Kalahari of Botswana.J . Arid Environ. 35 (3),563 - 574. Venier, L.A., Pearce,J. , McKee, J.E., McKenney, D.W., andNiemi , G.J., 2004. Climate and satellite-derived land cover for predicting breeding bird distribution in the Great Lakes Basin. J. Biogeogr. 31, 315-331. Vogelmann, J.E., Rock, B.N., 1989.Us e of thematic mapper data for the detection of forest damage caused by the pear thrips. Rem. Sen. Environ. 30 (3), 217-225. Ward, D.F., Lariviere, M-C, 2004. Terrestrial invertebrate surveys and rapid biodiversity assessment inNe w Zealand: lessons from Australia. New Zealand. J. Ecol. 28 (1), 151-159. White, D.,Minotti , P.G., Barczak, M.J., Sifneos, J.C., Freemark, K.E., Santelmann, M.V., Steinitz, M.J., Kiester, A.R., Preston, E.M., 1997.Assessin g risks to biodiversity from future landscape change. Conserv. Biol. 11(2) ,349-360 . Wolfgang, B.,2003 . Biodiversity and agri-environmental indicators—general scopes and skills with special reference to the habitat level. Agriculture, Ecosystems and Environment 98,35-78 . Zinner, D., Pelaez, F., Torkler, F.,2001 . Distribution and Habitat Associations of Baboons (Papio hamadryas) in Central Eritrea. Int. J. Primatology 22 (3),397-413 . E.J. Bell and R.C. Hinoja, Markov analysis of land use change: continuous time and stationary processes. Socio-Econ. Plan. Sci. 11(1977) , 13-17. 171 •V 4 %« •/' 0*41m A \k &•'«£• %• .as Chapter 7 Synthesis: Spatial complexity in the diversity puzzle: implications for conservation planning Euridice Leyequien Abarca Introduction To understand the environmental and controls on species diversity, we need to explore the relative differences of species richness patterns over avariet y of spatial scales. Current debates about patterns and processes regulating species diversity, do not explicitly consider the spatial scale (Huston 1999,Olf f etal. 2002). Previously, two levels of spatial scales were distinguished, i.e.; local and regional (Ricklefs and Schluter 1993). Subsequently, a hierarchical framework explaining the processes influencing species diversity has emerged (Table 1)(Willi s and Whirtaker2002) . In fact, itha s been claimed that the impact of processes atdifferen t spatial and/or temporal scales is central tounderstandin g global variation in biodiversity (Gaston 2000).However , studies at the species assemblage level usually focus only on a single spatial scale (e.g.;McGariga l and McComb 1995,Trzcinsk i etal. 1999,Hollan d and Fahrig 2000). It isplausibl e that species richness and abundance patterns respond differently to their environment at various scales (e.g.;Rolan d and Taylor 1997, Clarke and Lidgard 2000, Crawley and Harral 2001, Lennon etal. 2001 ,Rahbe k and Graves 2001, Leyequien etal. submitted). Often there is limited knowledge about the scales at which species respond to environmental factors, and this uncertainty may limit the effectiveness of study designs, and the applicability of results (Holland etal. 2004) . Thus, untangling the relative contribution of environment-species relationships in shaping ecological attributes at different spatial scales, such as species richness and abundance may provide valuable insights for solving the diversity puzzle (Meentemeyer and Box 1987, Wiens 1989,Levi n 1992, 1993,Borcar d etal. 1992,Dal e etal. 1994,Bissonett e 1997, Cushman and McGarigal 2002). 175 Table 1.Processes influencingspecies diversity ina hierarchicalframework across spatial scales (modifiedfrom: Willis and Whittaker2002). Spatial scale Scale of species Processes Important environmental richness variables responsible Local scale Species richness Resource partitioning, Resource availability, within communities, competition, predation, interacting species, physical within habitat patches facilitation habitat, precipitation, disturbance (e.g.; fire) Landscape scale Species richness Dispersal, migration, Topography, landscape between communities; colonisation/extinction, complexity, peninsula effect turnover of species (Metacommunity dynamics) within the landscape Regional scale Species richness of Migration/extinction Radiation budget and water large geographical dynamics, regional availability, area, latitude areas within constraints on persistence continents Continental Differences in species Speciation, colonisation, Acidification events, mountain- scale lineages and richness extinctions, biogeographical building episodes, e.g.; tertiary across continents history uplift of the Andes Global Differences reflected Biogeographical hitstory, Latitude, climate, inth e biogeographical productivity glacial/interglacial cycles ofth e realms, e.g.; Quaternary, continental plate distribution of movements, sea level change mammal families between continents In this research, we studied how biological and physical factors influence species diversity across different spatial scales in aNeotropica l bird community (Chapter 5). The results showed the responses ofbir d species richness and abundance tobiophysica l variables can be explained at a landscape scale. This isespeciall y the case for different migratory species such aswarbler s (e.g.; Wilsoniapusilla, Wilsonia citrina, Wilsonia canadensis, Dendroica townsendi, Dendroica virens andParula americana), and forest dependant species (e.g.; Aulacorhyncusprasinus, Lepidocolaptes ajfinis, Trogon elegans). Despite the fact that the landscape scale explained most ofth e variation in the species data, it is important to consider that the relationships between species richness/abundance and environmental variables vary at different spatial scales. In particular, there may be confounding effects between species richness and environmental variables,. Ake y result from this thesis is that appropriate spatial scales should be selected when studying how species respond to environmental variables. Properly identified scale is crucial for scientific studies or conservation planning, as scale is functionally significant for the target species (Grand and 176 Cushman 2003). The multi-scale approach presented in Chapter 5 could, besides the interest ofth e effect on community level,b e useful for species-specific analysis, for example for studies onNeartic-Neotropica l migrant species that currently show evidence of decline (e.g.; in this research: Chaeturavauxi, Contopuscooperi, Empidonax traillii, Empidonax hammondii, Empidonax wrightii,Empidonax difficilis, Oporornis tolmiei) (Seavey 2002); or for species falling under the Mexican category of special protection (e.g.; in this research: hawks,Accipiter striatus andButeo platypterus; toucans, Aulacorhyncus prasinus; woodpeckers, Campephilusguatemalensis; orvireos ,Hylophilus ochraceiceps) or endangered species (e.g.; in this research: solitaires,Myadestes unicolor; thrushes, Catharusfrantzii; quail-doves , Geotrygon albifacies;o rparrots ,Pionus senilis). Processes such ashabita t loss and fragmentation are relevant in understanding the differences in species diversity and are crucial factors affecting species diversity at landscape scale that could threaten species survival (Fahrig 1998,2003 ,Rosenzwei g 1995). How these inherently spatial processes affect species diversity across different levels, i.e.; low, intermediate or high levels of habitat loss and fragmentation may have important consequences for conservation planning. As concluded by Fahrig (2003),mos t researchers view habitat fragmentation as aproces s involving both the loss of habitat andth e breaking apart ofhabitat , and this has produced inconsistencies and difficulties in the interpretation of the effects ofbot h spatial processes on biodiversity. If conservation efforts aret o be successful, particularly when considering species assemblages, the individual effect of habitat loss and fragmentation on species richness needs to be elucidated. In our research we found that, when separating the individual effects of fragmentation and habitat loss,th e bird species richness show aunimoda l response to habitat fragmentation at landscape level with highest levels of species richness at intermediate fragmentation levels,wherea s a negative linear response to habitat loss (Chapter 3). Therelativ e importance ofbot h spatial processes did not prove significantly different in explaining species richness, and the interaction effect ofthes e two singleprocesse s was not significant. These results emphasise that both, habitat loss and fragmentation, are important factors influencing bird species diversity and that the individual effects of habitat loss do not outweigh the effects of 177 fragmentation on bird species richness. Wepropos e that conservation efforts should focus on both spatial processes together rather than on the factors separately, and we suggest that influencing fragmentation levels may mitigateth e negative effects of habitat loss on species richness (i.e.;th e spatial arrangement of habitat). At local scales, fragmentation of resources has frequently been shown to be associated with higher species richness (Tilman, 1982;Chaneto n and Facelli, 1991; Huston, 1994;Knap p etal, 1999),becaus e such differentiation prevents exclusion by a single superior competitor (Olff and Ritchie 2002). These results have important conservation implications ifth e design ofth e landscape matrix is set out to optimise bird species richness in aparticula r area. However, special attention should be drawn to the deterioration of natural habitat as aresul t of habitat fragmentation as itma y produce unsuitable conditions for species and therefore the chances for local coexistence of species can be counteracted (Olff and Ritchie 2002), or if fragmentation negatively affect the balance between the colonization and extinction rate of suitable habitat patches as itha s been found in a regional scale (Hanski and Simberloff 1997). Thediversit y in species assemblages can beregulate d by factors such as interspecific competition, and great attention had been devoted to understand the role of competition in the structure and dynamics of ecological communities (Ricklefs 1975, Giller 1984, Wiens 1989, Keddy 2001). Despite the large amount of research on the mechanisms of interspecific competition, the impact of competition on community composition has remained unclear. Similarity in body mass is an easily determined animal characteristic that probably could serve as aprox y for competition strength in ecological communities. However, the impact of a large overlap in body mass in structuring ecological communities remains unproven, and there is an ongoing debate about the relationship between species coexistence and body size differences among ecologically similar species. The thresholds for body size differences above which competition isminimized , are still not understood. In thisthesi s we addressed how interspecific competition is influenced by differences in body mass among ecologically comparable species (Chapter 4). The results showed that there is a significant negative relationship between the body mass ratio and the strength of 178 competition, suggesting that a large difference inbod y sizes among sympatric species promotes coexistence in communities. The species pairs range from relatively high body masses like granivorous birds like parrots or oropendola, i.e.;Pionus senilis-Psarocolius montezuma,t o small body masses like hummingbirds, e.g.;Amazilia candida- Campylopterus curvipennis (nectarivorous). This allometric relationship, though not causal, proves that the body mass ratio influences the strength of competition, and possibly ultimately the community species composition. Apart from understanding the factors controlling species diversity, andho w spatial patterns and processes at various spatial scales influence species richness and community patterns, it iscrucia l for conservation planners to efficiently monitor and assess trends in ecological communities. Specifically, in avian studies the selection of different techniques (e.g.; point counts and mist netting) to sample bird species richness and abundances may have profound consequences for the reliability and completeness ofth e study, and subsequently for the knowledge on the patterns in avian communities. The evaluation performed in Chapter 2 demonstrated that even after a large sampling effort, point counts perform better in detecting species richness with a lower total effort and costs, than mist netting. However, we suggest thatwhe n possible both techniques should be usedt o benefit from their individual strengths (e.g.;detectio n of understory, canopy species or secretive species). Therelativel y higher costs of mist netting compared topoin t counts,i fw ewoul d have aimed atrecordin g 90 % ofth e species,woul d be 3297 USD,whic h is 2090 USD (calculated for the period 2033-2004) or almost twice total costs for point counts. Nonetheless, the higher costs of mist netting could bebalance d by the higher efficiency in registering for example understory species that were missed by point counts (i.e.;Mionectes oleagineus, Catharusaurantiirostris, Catharusmexicanus, Helmitheros vermivorus, Oporornisformosus, Seiurus motacilla,Basileuterus belli, Melospiza lincolnii, Cyanocompsaparellina, Passerina cyanea andPasserina ciris), butthi s depends on the research objectives (Chapter 2).Furthermore , the generation of consistent and reliable information about species distribution and diversity is critical inbiodiversit y conservation (Riede 2000). One promising approach for the assessment, monitoring and prediction of 179 animal diversity is the synergy of remote sensed information and auxiliary data with ecological models, and a subsequent validation using traditional field techniques (Chapter 6). An example ofthi s isth e relatively recent development of methods using remotely sensed data in ecological species distribution models (see Chapter 6 for areview) . However, despite the potential ofremot e sensing applications in ecology, there is still a considerable time lag time between the developments of new remote sensing techniques and their use in ecological research. Conservation implications: shade coffee plantations as a refuge for Neotropical birds Biogeographicalposition of thestudy area: its importancefor conservation The extended region of the study area of this thesis, the northeastern mountain range of Puebla, is located in a strategic zone alongth e border ofth eNeotropica l region and the Nearctic region (Rzedowski 1966). It is situated in the meridional extreme ofth e Sierra Madre Oriental and borders two major landscapes: the Mexican Transvolcanic Belt ('Eje neovolcanico') and the Coastal lowlands. The study area covers the vegetation belts of lower montane rain forest and lowland rain forest (both tropical evergreen forests). This geographic position of the northeastern mountain range of Puebla isa key biogeographical location characterized by high biodiversity. For example, there are 558 registered plant species in different tropical evergreen forests in SianKa'a n (Quintana Roo),73 7 in Tuxtepec (Oaxaca), 800 in Uxpanapa (Veracruz), 814 in Los Tuxtlas (Veracruz) and 984 in Montes Azules (Chiapas). For other taxa such as mammals the recorded species richness is 85 in Montes Azules (Chiapas) and 90 in Los Tuxtlas (Veracruz). In particular, avian diversity is high in the tropical evergreen forests in Mexico, for example, 150 species in Uxpanapa (Veracruz), 301 in Peten (Yucatan) and 341 in Los Tuxtlas (Veracruz). 180 The study area, aspar t ofth e northeastern mountain range of Puebla, is an important region for bird conservation and is also part of the migratory route for the Nearctic-Neotropical birds. Specifically, during the fieldwork for this thesis we documented atota l of 181bir d species within the traditional shade coffee plantations, representing atota l of 12orders ,3 1 families and 123gener a (Annex 1).Th e 181bir d species represent 61 % of the total bird species (298) registered for the northeastern mountain range of Puebla and 17% of all Mexican birds (1060 species).Thepredicte d species richness (Smax)fo r the study area shows that the asymptote was nearly reached (Figure 1),indicatin g that the sampling effort in this study enabled us to detect the far majority of allbir d species. From the 181 species, 124(6 9 %) were resident birds and 57 (31 %)Neartic-Neotropica l migrants, 9 (5 %) were forest- dependent species, 4 (2 %) endemics, 10( 5 %) were subject to special protection and 4 (2 %) were endangered (Norma Oficial Mexicana NOM-059-ECOL-2001). - - -Observe d Extrapolation effort — Predicted 1 10 19 28 37 46 55 64 73 82 91 10010 911 812 713 614 515 416 317 218 119 0 19920 821 722 623 524 425 3 Sampleday s Figure 1.Accumulation curve (totalobserved species richness of 182)andpredicted Sm (164.77), the light lineshows the extrapolation effort reaching an asymptote atS max of 165.24. 181 Despite the biogeographical and ecological importance of the northeastern mountain range of Puebla, deforestation has caused a serious loss of habitat for wildlife species in the last thirty years, mainly due to deforestation and conversion to agricultural lands and pasturelands (Lugo-Hubp etal. 2005) . In Mexico, annual deforestation rates for tropical evergreen forests are estimated at 1.9% for lower montane rain forest and 2 % for lowland rain forest. This is considerably higher than for temperate (upper montane) forests with 0.64 %/y, specifically for coniferous and broadleaved forests (Masera etal. 1997). Some of the major causes that threaten both theNeotropica l birds andth e migrants are habitat loss and fragmentation by wide-spread deforestation. In particular, important declines in Neotropical migrants havebee n found (Rappole 1995,Robbin s etal. 1995), and itha s been suggested that one ofth e major causes for these declines is caused byproblem s inthei r wintering areas (Estrada and Coates-Estrada 2005). Atpresent , an important agricultural land cover within the tropical evergreen forests is coffee, the majority of which belong to the traditional shade system. In Mexico, coffee- growing regions fall largely within areas identified asbiodiversit y hotspots. Evidence suggests that shade coffee agro-ecosystems have important ecological functions, such as biodiversity maintenance and protection (for a review seePerfect o andArmbrech t 2003), water, health and production, soil fertility, capture of carbon andrain ,refuge s for biological control agents and pollinators, and maintenance of climatic balance (Toledo and Moguel, 2006). Shade coffee plantations are also characterised by ahig h diversity and abundance of plants. A floristic survey carried out inth e region at 31 sites with traditional shade coffee plantations (Toledo etal. 2004) ,reporte d between 35 and 150plan t species per hectare, with an estimated total of 250 species (Figure 2).Thes e coffeegardens are man-made systems reflecting in their structure and composition the traditional management of plant species (e.g.;domesticate d and introduced species). In the northeastern mountain range of Puebla, 96%o fth e recorded species in the traditional shade coffee plantations are useful, with a clear dominance of species serving for food (almost 50%),followe d by species serving for medicines, materials for construction, fuelwood and others.A remarkable group are the edible fruits that also constitute food resources for birds: sapotes (14 182 varieties), avocados (8varieties) ,berrie s (17 species), citrus species (14 varieties) and others. Similarly, there are abundant plant taxa that constitute nectar sources for birds: 11 varieties ofplatanillo s (Heliconia spp.),4 species of Zingiberaceae, 3 species of Marantaceae and others. Thus, as habitats for the avifauna, coffee gardens are not only forest refuges, but alsoprovid e abundant food resources for birds (e.g.;nectar , fruit and insects) (Wunderle and Latta 1998;Greenber g etal. 1997a, b). Some examples of the most common plant species arePimento dioica,Inga lactibracteata andHamelia patens as source of insects and nectar, andBursera simaruba, Tremamicrantha, Psidium guajava andPersea americana as source of fruits. Thus, the maintenance of traditional shade coffee agroecosystems inthi s landscape mosaic can assist in the conservation of Neotropical avifauna, and help maintain multiple ecological services. 12 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Sample days Figure 2.Accumulation curve of recordedplant species at 31sites withshaded grown coffee in thenortheastern mountain range ofPuebla. Source: Toledo,et al. 2004. 183 Shade coffeeplantations and theanthropogenic matrix From the total research area (151,367.5 ha),56,36 4 ha (31 %) comprise agricultural lands (i.e.; non-arboreal crops) and 11,925 ha (7 %) are shade coffee plantations. In addition there are 15527h a (9 %) of tropical evergreen forest (i.e.; lowland rain forest and lower montane forest) and 67,552 ha (37 %) of a diversity of forest formations (i.e.;a t different succession and/or management stages), whilst the remaining area is composed by urban areas and water bodies.A s noted above shade coffee plantations represent aforeste d habitat for wildlife, including avifauna (i.e.; 124resident birds, 57 migratory birds and 9 forest dependent species in our research area),whic h isparticularl y important in areas with high degree of fragmentation and habitat loss.Accordin g to Perfecto and Armbrecht (2003), shade coffee agroecosystems could provide ahigh-qualit y matrix for many bird forest species orbir d species that originally required natural closed-canopy forests. In fact, we found that the total area covered by coffee plantations has a significant positive influence (p =0.05 ) on species richness (Chapter 3). This confirms other studies which have shown that the expected species richness (of migrant birds) in shade coffee plantations is significant higher than in other agricultural systems (Figure 3) (Estrada and Coates-Estrada 2005). The area of traditional shade coffee plantations as well as their ecological qualities constitute a key element inth e design of a landscape that aims at optimising biological conservation and providing human subsistence. We suggest that the conservation plans ofthi s region could incorporate the paradigm of reconciliation ecology (Rosenzweig 2005) or bioregional conservation (Toledo 2005) and adopt some the results found here. The value of protected areas for the preservation and maintenance ofbiologica l diversity is doubtful if the conservation strategies do not actively incorporate areas managed for human use (Bengtsson et al.2003) . 184 400 600 1000 Numbero f individuals Figure 3.Rarefaction curvesfor different agroecosystems: shadeplantations, unshade plantations, livefences, non-arboreal crops andpasture lands (source:Estrada and Coates-Estrada 2005). Traditional shade coffee plantations in the international context: scenarios for biodiversity conservation and carbon sequestration Biodiversity conservation and carbon sequestration are key ecological issues that are subject of a common international concern and numerous cooperation programmes. As has been mentioned in the introduction of this thesis the current trend in northern Latin America (Colombia, Panama, Costa Rica, El Salvador, Guatemala, Nicaragua, Honduras and Mexico) ist o reduce shade cover plantations and convert them into unshaded coffee plantations or other open non-arboreal agricultural systems. If this trend continues it could have important implications for the loss of biodiversity including bird species, and also loss of important ecological services that the shade coffee plantations provide such as carbon sequestration. Here we present scenarios of the potential consequences on: a) species richness (specifically avifauna) and b) carbon sequestration, ifth e area of shade coffee plantations continues to be reduced. 185 a) Species richness. The Convention on Biological Diversity in Rio de Janeiro in 1992 is the result of international cooperation to anticipate, prevent and battle the causes of a significant reduction or loss ofbiologica l diversity. Traditional shade coffee plantations support a high biodiversity compared to open sun coffee plantations or other non-arboreal open agricultural systems, andtherefor e they represent a suitable habitat for awid e range of species. Specifically in Latin America, the bird diversity had been studied in shade coffee plantations andth e results show arelativel y high avian diversity ofresident birds and also migrants (Table 2) (Aguilar-Ortiz 1982, Robbins etal. 1992, Vennini 1994,Wunderl e and Latta 1996, Greenberg et al. 1997, Van der Voort and Greenberg 1999, Dietsch 2000, Johnson 2000). Therefore, the loss of this man made habitat (i.e.;traditiona l shade coffee plantations) could have a deleterious effect on the bird species richness. Ifw e take the example of the resident birds in our research area with an observed species richness of 124an d Smax of 113.59(Figur e 4),w e found that based on our analyses (Chapter 3) using the extended species-area curve, we can expect an estimated decline of 0.006 in species richness for every hectare the coffee area is reduced. Theseresult s could have significant consequences for conservation strategies asth e amount of traditional shade coffee plantations has apositiv e linear relationship with the species richness of resident birds (p<0.001). 186 Table2. Bird speciesrichness found in tropicalforest, shade coffeeplantations, rustic coffee and unshaded coffee S, Species richness; n.a.,Not Available. Habitat Observed S Estimated S Country Country Tropical forest 75 n.a. Mexico Greenberg e?a/ . (1997b) 86 77 Mexico Tejeda-Cruz and Sutherland (2004) 131 n.a. Panama Petite/al. (1999) 82 81.6 Mexico Greenberg etal. (1997a ) 85 78 Mexico Tejeda-Cruz and Sutherland (2004) 138 n.a. Mexico Aguilar-Ortiz(1981) 96 n.a. Mexico Navarro (1992) 51 46 Mexico Tejeda-Cruz and Sutherland (2004) Shade coffee 56 55.5 Mexico Greenberg etal. (1997) 136 n.a. Mexico Aguilar-Ortiz(1981) 65 64.7 Guatemala Greenberg etal. (1997b) 90 n.a. Mexico Calvoan d Blake (1998) 41 n.a. Mexico Johnson(2000) 87 n.a. Guatemala Petit era/. (1999) 79 69 Jamaica Tejeda-Cruz and Sutherland (2004) 65 n.a. Mexico Cruz-Angon and Greenberg (2005) 99 n.a. Mexico Arangon and Paniagua (1989) 182 164.77 Mexico Leyequien (2006) 99 n.a. Mexico Borrero(1989) Rustic coffee 69 68.7 Mexico Greenberg etal (1997b) 80 70 Mexico Tejeda-Cruz and Sutherland (2004) Unshaded 49 49.2 Mexico Greenberg etal (1997b) 75 n.a. Mexico Calvo and Blake (1998) 6 4 Guatemala Tejeda-Cruz and Sutherland (2004) 20 40 80 100 Sample days Figure 4. Accumulation curvefor resident birds (observedspecies richness of 125)and predictedSmax (113.59). 187 b) Carbonsequestration. Under the Kyoto Protocol, several countries agreed to adopt specific targets for a reduction in greenhouse gas emissions within the first Kyoto commitment period (2008-2012), and carbon credits in the Certified Emission Reductions market have been offered to reduce the C02 emissions with an actual market value ofUSD $ 5/tC02eq (Teixeira-Coelho etal. 2005). One form of emissions offset isth e capture of carbon in carbon sinks. Carbon sequestration involves the transfer of atmospheric C02 into securelong-live d pools thus preventing the immediate reemission to the atmosphere. Therefore, throughjudiciou s land use and management practices, soil carbon sequestration can increase the soil organic carbon and the soil inorganic carbon stocks (Lai 2004). Agroforestry systems have ahig h and attainable soil carbon sequestration potential with arat e of 100-200k g C/ha per year compared to extractive farming practices (Lai 2004). Traditional shade coffee plantations fall in the category of agroforestry, and therefore have clear benefits for carbon sequestration (Kursten and Burschel 1993,Schroede r 1994,D e Jong etal. 1997,Kurste n 2000,Avil a etal. 2001,Noordwij k etal. 2002 , Pandey 2002, Suarez-Pascua 2002,Albrech t and Kandji 2003), moreover traditional shade coffee agroecosystems add high amounts of biomass to the soil and prevent soil loss,conserv e water (Beer etal. 1998) and enhance activity and species diversity of soil fauna (Nestel 1995). Hence, carbon-trading markets offering C-credits may be a solution to mitigate the negative impacts of climate change and other ecological problems (e.g.; soil degradation and disturbance in hydrological cycles). The calculated carbon stock by the area under traditional shade coffee plantations in Mexico (70 % of the total coffee area) is 11,255,973t C, based on an average carbon stock of21.61 C ha"1 ofth e total aerial C-stock including shade trees, coffee bushes and leaf litter (derived from Alpizar etal. 1985,Avila-Vargas 2000, Suarez-Pascua 2002,Potzo l 2004). This represents approximately 56,279,864 USD in carbon credits (USD5/1C). However, the deforestation in shade coffee plantations has been estimated at arat e of 0.4 %pe r year (Blackman etal. 2005)Wit h this rate of deforestation in twenty years there will be a reduction of 89% of carbon stock if it converts intopastur e lands or extractive 188 farming. In our research area the calculated carbon stock ofth e total area covered by shade coffee plantations is257,7891 C based on the same average carbon stock of 21.6 t C ha"1,representin g 1,288,943US D in carbon credits. Moreover, ifth e conversion rate would be 0.4 %th e results in carbon loss will be devastating injus t one year. However, the conversion rate for the research area isunknow n and we approximate an almost negligible rate of conversion compared to the previously reported. The potential for the coffee producers as suppliers of carbon sequestration is calculated in approximately 105USD/ha , which represents a considerable income compared to approximately an average of 59USD/h a per coffee harvest. Traditional shade coffee plantations in the regional context: Adaptative management byindigenou s people In Mexico, the coffee regions overlap largely with areas supporting a high biological richness (Moguel and Toledo 1999). It is estimated that 60-70 % of coffee plantations in Mexico are shade coffee plantations (Moguel and Toledo inprep). Traditional shade coffee plantations are cultivated mainly by small-scale community-based growers, the majority of them belong to some indigenous group (Moguel and Toledo 1999). The indigenous mode of utilisation of this tropical agroecosystems denotes a multiple use strategy (MUS), which can be called adaptative management (Toledo et al. 2003). This indigenous adaptative management has obvious socio-economic as well as ecological benefits. The ecological advantages of shade coffee agroecosystems are clear: a) high biodiversity maintenance, b) regulation of carbon cycle, c) soil protection, d) regulation of hydrological cycle and e) preservation of forest cover. The economic benefits derived from this tropical agroecosystems are sources of goods, services and energy for household subsistence and products for local, regional and international markets. For example, shaded coffee plantations hold ahig h proportion of useful plant diversity. Estimates of useful plants are in a range of 35 to 150 species per hectare (mainly tropical fruits, spices, ornamentals, bamboos and medicinal plants) (Moguel and Toledo inprep.) This plant species richness has a substantial potential for international markets oriented to organic agricultural products 189 as well as for conventional products. Indigenous multiple use management creates a diverse landscape mosaic which could be approached in a three dimensional view in space, in its horizontal axis as multiple land covers at different stages of use (e.g.; the use of succession stages of vegetation for agriculture), in a vertical axis as the structural complexity of the MUS (e.g.; vegetation structural complexity) and the third axis as the biological diversity and composition resulted from the MUS. Maintaining such a functional and productive landscape with diverse land uses and ecosystems seems to be an optimal strategy to maintain biodiversity (Pimentel et al. 1992, Toledo etal. 2003) . Moreover, since the coffee crisis in the last decade as a result of overproduction in Vietnam and Brazil (ICO 2003), and the conversion of shade plantations into unshaded plantations in countries like Colombia or Costa Rica (Perfecto et al. 1996) new market alternatives as for example the "sustainable coffee" or the "shade coffee" certification programs emerge as a way to improve the economic situation of coffee producers in Latin America together with the maintenance of biodiversity (Perfecto et al. 2005). Several studies had approached the problem between yield and biodiversity maintenance in which the construction of a yield set from functions of shade cover proved to be successful for optimising both yield and biodiversity (Staver et al. 2001,Perfect o et al. 2005). Perfecto et al. (2005) concluded already that there are two critical factors for decision making in these certification programmes: a) the relationship between shade and yield is not linear (it has been found in many studies that this relation is best described by a humped-shape curve) (Soto-Pinto et al. 2000, Staver et al. 2001), and b) different taxa or species could differ in their sensitivity to changes in the structural and floristic composition of vegetation (i.e.; reduce of shade trees). Another feasible market option for coffee farmers is the transformation from conventional to organic coffee, which represents an increase of yield ranging from 15 % (Bray et al. 2001) to 67 % (Damiani 2002). Thus, the balance between yield and the maintenance of biodiversity needs to be accounted if successful shade coffee certifications programs are to be implemented and most likely this process should involve producers aswel l as government and non-government institutions. 190 In the research area of this thesis, the main managers of the shade coffee plantations are indigenous people from two groups, i.e.; nahuats and totonacos. One of the major indigenous organisations in the northeastern mountain range of Puebla is Tosepan Tatataniske, which has approximately 30,000 members corresponding to 60 % of the regional total population (Moguel and Toledo in prep). Currently they manage approximately 5,800 ha of shade coffee plantations, which play a crucial role in the maintenance of the forested cover in the landscape along with the associated ecological services. Furthermore, the ecological knowledge on the regional avifauna enables these indigenous groups to influence the structural and floristic design of the shade coffee plantations to a certain extent, where food resources for birds are explicitly maintained in the plantations (Leyequien et al. inprep.). Coffee farmers identified plant species within their coffee plantations that are food resources for a variety of birds, mainly maintained in the plantation as sources of cash crops or self-subsistence products, but also to attract birds for mere contentment. This ecological indigenous knowledge acquired over a long history of natural resources utilisation, should be considered in the conservation strategies if we successfully want to maintain and preserve the biological diversity within the complexity of the anthropogenic matrix. 191 References Aguilar-Ortiz, F. 1982.Estudio ecologicod e las aves del cafetal, in Estudios Ecologicos en el Agroecosistema Cafetalero, Jimenez-Avila, E and Gomez-Pompa,A . (eds.) INIREB,Xalapa , Veracruz, Mexico. Albrecht, A. and Kandji, S.T. 2003. Carbon sequestration intropica l agroforestry systems. Review. Agriculture, Ecosystems and Environment 99: 15-27. Alpizar, L., Fassbender, H.W. and Heuvelop,J . 1985.Sistema s agroforestales de cafe (Coffeaarabicd) con laurel (Cordiaalliodord) y con Poro (Erythrina poeppigiana) en Turrialba, Costa Rica. I. Biomasa y reservas nutritivas. Turrialba 35(3): 233-242. Avila, G., Jimenez, F.,Beer , J., Gomez, M. and Ibrahim, M. 2001. Almacenamiento, fijacion de carbono y valoracion de servicios ambientales en sistemas agroforestales en Costa Rica. Agroforesteria en lasAmerica s 8(30):32-35 . Avila Vargas, G. 2000. Fijacion y Almacenamiento de Carbono en Sistemas de Cafe Bajo Sombra, Cafe a Pleno Sol, Sistemas Silvopastoriles yPastura s a Pleno Sol. Turrialba, Costa Rica: CATIEthesis . Beer, J., Muschler, R., Kass, D. and Somarriba, E.1998. Shade management in coffee and cacao plantations. Agroforestry Systems 38: 139-164. Bengtsson, J., Angelstam, P., Elmqvist, T., Emanuelsson, U., Folke, C, Ihse, M., Moberg, F. and Nystrom, M. 2003. Reserves,resilienc e and dynamics landscapes. Ambio 32(6): 389-396. Bissonette, J.A., Harrison, D.J., Hargis, CD. and Chapin, T.G. 1997.Th e influence of spatial scale and scale sensitive properties in habitat selection by American Martins. In: Bissonette, J.A. (ed.) Wildlife and landscape ecology: effects of patterns and scale. Springer-Verlag, New York. Blackman, A., Albers, H., Avalos-Sartorio, B.An d Crooks, L. 2005.Deforestatio n and shade coffee in Oaxaca, Mexico. Key research findings. RFF DP 05-39. Washington, D.C. 192 Borcard, D., Legendere, P. and Drapeau, P. 1992.Partiallin g out the spatial component of ecological variation. Ecology 73(3):1045-1055. Borrero, H. 1986.L a substitution de cafetales de sombriopo r caturrales y su efecto negativo sobre la fauna de vertebrados, Caldasia, 15:725-7326 . Bray, D.B.,Plaz a Sanchez, J.L. and Murphy, E.C. 2001.Socia l dimensions of organic coffee production in Mexico: lessons for eco-labeling initiatives. Soc.Nat . Resour. 15:429-446. Clarke, A. and Lidgard, S.J. 2000. Spatial patterns of diversity in the sea: bryozoan species richness in the north Atlantic. J. Anim. Ecol.69: 799-814. Crawley, MJ. and Harral,J.E . 2001.Scal e dependency inplan t biodiversity. Science291 : 864-868. Cushman S.A. and McGarigal, K. 2002.Hierarchical , multiscale decomposition of species- environment relationships. Landscape Ecology 17:637-646. Chaneton, E.J. and Facelli,J.M . 1991. Disturbance effects on plant community diversity: spatial scales and dominance hierarchies. Vegetatio 93: 143-156. Dale, V.H., Pearson, S.M., Offerman, H.L. and O'Neill, R.V. 1994.Relatin g patterns of land use change to faunal biodiversity in the central Amazon. Conserv Biol 8:1027-1036. Damiani, O.2002 . Organic Agriculture in Guatemala: A Study of Producer Organizations in the Cuchumatanes Highlands. Office of Evaluation and Studies, IFAD , Rome. De Jong,B.H . etal. 1997. Forestry and agroforestry land-use systems for carbon mitigation: an analysis in Chiapas, Mexico. In: Adger, W.N., Whitby, M., and Pettenella, D.M. (eds.) Climate-Change Mitigation and European Land-Use Policies, CAB International, New York. Dietsch, T.V. 2000. Assessing the conservation value of shade-grown coffee: a biological perspective using neotropical birds.Endangere d Species Update 1750-72 [Electronic version].Retrieve d January 2, 2006, from http://www.umich.edu/- esupdatel 193 Donald, P.,2004 . Biodiversity Impacts of Some Agricultural Commodity Production Systems. Conservation Biology 18:17-37. Estrada, A. and Coates-Estrada, R. 2005. Diversity ofNeotropica l migratory landbird species assemblages in forest fragments and man-made vegetation in Los Tuxtlas, Mexico. Biodiversity and Conservation 14: 1719-1734. Fahrig L. 1998.Whe n does fragmentation of breeding habitat affect population survival? Ecol. Model. 105:273-92 Fahrig, L. 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Syst. 34:487-515. FAO 1996.Fores t Resources Assessment 1990:Surve y of Tropical Forest Cover and Study of Change Processes. Rome, Italy: FAO, FAO Forestry Paper 130: 152pp . FRA 2005.Globa l forest resources assessment update 2005. Specification of national reporting tables for FRA 2005. Rome, Italy: FAO, FAO Forestry Paper 81: 37pp . Gaston, K.J. and Blackburn, T.M. 2000. Patterns and process in macroecology. Blackwell. Giller, P.S. 1984. Community structure and the niche. Chapman and Hall, London. Grand, J. and Cushman, S.A. 2003. A multi-scale analysis of species-environment relationships: breeding birds in apitc h pine-scrub oak (Pinus rigida-Quercus ilicifolia)community . Biol. Conserv. 112:301-317. Greenberg, R., Bichier P. and Sterling J. 1997. Bird populations in rustic and planted shade coffee plantations ofEaster n Chiapas, Mexico. Biotropica 29: 501-514. Hanski, I. and Simberloff, D. 1997. The metapopulation approach, its history, conceptual domain and application to conservation. In:Hanski , I. and Gilpin, M.E. (eds.) Metapopulation Biology. Ecology, Genetics and Evolution. Academic Press, San Diego. Holland, J.D. and Fahrig, L. 2000. Effect of woody borders on insect density and diversity in crop fields: A landscape-scale analysis. Agriculture, Ecosystems and Environment 78: 115-122. Holland, J.D., Bert, D.G. and Fahrig, L. 2004. Determining the spatial scale of species' responses to habitat. BioScience 54(3):227-233 . 194 Huston, M.A. 1999. Local processes and regional patterns: appropriate scales for understanding variation inth e diversity ofplant s and animals. Oikos 86:393—401 . Huston, M.A. 1994.Biologica l diversity. The Coexistence of Species on Changing Landscapes. Cambridge University Press, Cambridge. ICO 2003. Website ofth e International Coffee Organisation: http://www.ico.org Johnson, M.D. 2000. Effects of shade-tree species and crop structure on the winter arthropod and bird communities in aJamaica n shade coffee plantation. Biotropica 32: 133-145. Keddy, P.A. 2001.Competition . Kluwer Academic Publishers, Dordrecht. Knapp,A.K. ,Blair , J.M., Briggs,J.M. , Collins, S.L., Hartnett, D.C.,Johnson , L.C. and Towne, E.G. 1999. The keystone role of bison in North American tall grass prairie: bison increase habitat heterogeneity and alter abroa d array of plant, community, and ecosystem processes. Bioscience 49, 39-50. Kursten, E. 2000. Fuelwood production in agroforestry systems for sustainaible land use and C02-mitigation. Ecological Engineering 16:S69-S72 . S? Kursten, E. and Burschel, P. 1993.C02-mitigatio n by agroforestry. Water, air and soil 70: 533-544. Lai, R. 2004. Soil carbon sequestration impacts on global climate change and food security. Science 304: 1623-1627. Lennon, J.J.,Koleff , P.,Greenwood , J.J.D. and Gaston, K.J. 2001. The geographical structure of British bird distributions: diversity, spatial turnover and scale. J. Anim. Ecol. 70: 966-979. Levin, S.A. 1992. The problem of pattern and scale in ecology. Ecology 73:1943-1967. Levin, S.A. 1993.Concept s of scale atth e local level. In: J. R. Ehleringer and Field, C.B . (eds.) Scaling physiological processes: leaf to globe. Academic Press,Ne w York, New York, USA. Leyequien etal. {in prep) Etno-ornithology in the northeastern mountain range of Puebla. Leyequien, E., De Boer, W.F. and Skidmore, A.K. (submitted) Linking species- environment relationships and multiple spatial scales in community ecology. 195 Lugo-Hubp, J.,Zamorano-Orozco , J.J.,Capra , L.,Invar. , M.an dAlcantara-Ayala , I.2005 . Revista Mexicana deCiencia s Geologicas 22(2):212-228 . Masera, O.R.,Ordones , M.J. andDirzo , R.1997. Carbon emissions from Mexican forests: current situation andlong-ter m scenarios. Climatic Change 35:265-295 . Mcgarigal, K.an dMcComb , W.C. 1995.Relationship s between landscape structurean d breeding birds inth eOrego n Coast Range. Ecol. Monogr. 65:235-260 . Meentemeyer, V. andBox , E.O. 1987. Scale effects inlandscap e studies.In :Turner ,M.G. (ed.) Landscape heterogeneity anddisturbance . Springer-Verlag,Ne wYork , USA. Moguel,P .an dToledo , V.M. 1999. Biodiversity conservation intraditiona l coffee systems in Mexico. Conservation Biology 13: 1-12. Moguel, P.an dToledo ,V.M . {inprep) Indigenous resistance, shaded coffee and globalization: the"Tosepa n Titataniske" Cooperative of Mexico. Nestel, D. 1995.Coffee inMexico : international market, agricultural landscape andecology . Ecol. Econ.15: 165-178. NOM-059-ECOL. 2001.Environmenta l protection—native species from theflor a and fauna of Mexico—risk categories andguideline s forthei r consideration inth elis t of species indange r of extinction, threatened orprotected . List ofspecie si n danger ofextinction . Official Paper ofth e Mexican Federation, March 6,2002 , Mexico (in Spanish). Olff, H.an dRitchie , M.E. 2000. Fragmented nature: Consequences forbiodiversity . 3rd International Workshop onSustainabl e Land Use Planning, Wageningen,th e Netherlands. Olff, H.,Ritchie , M.E .an dPrins , H.T .2002 . Global environmental controls ofdiversit y in large herbivores. Nature 415: 901-904. Pandey, D.N. 2002. Carbon sequestration inagroforestr y systems. Climate Change Policy 2: 367-377. Perfecto, I.,Rice ,R. ,Greenberg , R.an dVa nde rVoort , M. 1996. Shade coffee: a disappearing refuge forbiodiversity . BioScience 46(8):598-603 . 196 Perfecto, I. and Armbrecht, I. 2003. The coffee agroecosystem in theNeotropics : combining ecological and economic goals. In: Vandermeer, J. (ed.) Tropical agroecosystems. CRC Press, Washington, D.C. Perfecto, I., Vandermeer, J., Mas,A . And Soto-Pinto, L. 2005.Ecologica l Economics 54: 435-446. Pimentel, D., Stachow, U., Takacs, D.A. and Brubaker, H.W. 1992. Conserving biological diversity in agricultural/forestry systems. Bioscience 42:354-362 . Potzol, C. 2004. Carbon storage in coffee agroecosystems of southern Costa Rica: potential applications for the Clean development Mechanism. MSc thesis.Yor k University, Toronto, Ontario, Canada. Rahbek, C. and Graves, G.R. 2001. Multiscale assessment ofpattern s of avian species richness. Proc.Natl . Acad. U.S.A. 98(8):4534-4539 . Rappole J.H. 1995. The Ecology of Migrant Birds. Smithsonian Institution Press, Washington, DC. Ricklefs 1975. Patterns of growth inbirds , part 3,Growt h and development ofth e Cactus Wren. Condor 77:34-45 . Ricklefs, R.E. and Schluter, D. 1993.Specie s diversity: regional and historical influences. In: Ricklefs, R.E., Schluter, D. (Eds.), Species Diversity in Ecological Communities. University of Chicago Press, Chicago. Riede, K. 2000. Conservation and modern information technologies: the global register of migratory species (GROMS). J. of Int.Wildlif e Law and Policy 3(2): 152-164. Robbins, C.S. Dowell, B.A. , Dawson, D. K., Colon, J. A., Estrada, R., Sutton, A., Sutton, R. and Weyer, D. 1992. Comparison ofNeotropica l migrant landbird populations wintering in tropical forests, isolated forest fragments, and agricultural habitats. In: Ecology and Conservation ofNeotropica l Migrant Landbirds, Hagan, J.M. and Johnson, D.W., Eds., Smithsonian Institution Press, Washington, D.C. Robbins, C.S., Dawson, D.K. and Dowell, B.A. 1995.Changin g land use problems and opportunities. In: Wilson, M. and Sader, S. (eds.), Conservation of Neotropical Migratory Birds in Mexico. Maine Agricultural and Forest Experiment Station Miscellaneous Publication 727,pp . 1-14. 197 Roland, J. and Taylor, P.D. 1997. Insect parasitoid species respond to forest structure at different spatial scales.Natur e 386:710-713 . Rosenzweig, M.L. 1995.Species diversity inspace and time. Cambridge University Press, Cambridge, united Kingdom. Rosenzweig, M.L. 2005.Avoidin g mass extinction: Basic and applied challenges.24 th Annual Midwest Ecology and evolution Conference. Dept. of ecology and Evolutionary Biology, University of Arizona, Tucson. Rzedowski,J . 1966Vegetacio n del estado de San Luis Potosi.Act a Cientifica Potosina 5: 5-291. Rzedowski, J. 1978.Vegetacio n deMexico .E d Limusa.Mexico , D.F . Schroeder, P. 1994. Carbon storage benefits of agroforestry systems. Agroforestry Systems 27: 89-97. Soto-Pinto, L.,Perfecto , I., Castillo-Hernandez, J., Caballero-Nieto,J . 2000. Shade effect on coffee production atth enorther n Tzeltal Zone ofth e state of Chiapas, Mexico. Agric, Ecosyst. and Environ. 80:61-69. Staver, C, Guharay, F.,Monterroso , D. and Muschler, R.G. 2001.Designin g pest suppressive multistrata perennial crop systems: shade grown coffee in Central America. Agrofor. Syst. 53: 151-170 . Suarez Pascua, D. 2002. Cuantificacion y Valoracion Economica del Servicio Ambiental Almacenamiento de Carbono en Sistemas Agroforestales de Cafe en la Comarca Yassica Sur, Metagalpa, Nicaragua. Turrialba, Costa Rica: CATIE thesis Teixeira-Coelho, S., Moreira, J.R., Martins,A.S. ,Gonzalez-Velazques , S.M.S. and Castelo-Branco, M.B.2005 .Galler y forest recovery and carbon credits. RIO 5- World Climate and Energy Event, Rio de Janeiro, Brazil. Tilman,D . 1982.Resourc e Competition and Community Structure. Princeton University Press, Princeton. Toledo, V.M. 2005.Repensa r la conservation: ^,areasnaturales protegida s o estrategia bioregional? Gaceta ecologica 77:67-82 . Toledo, V.M., Ortiz-Espejel, B., Cortes, L.,Moguel , P. and Ordonez etal. 2003 . The multiple use of tropical forests by indigenous peoples inMexico :cas e of 198 adaptative management. Conservation Ecology 7(3): 9. [online] URL: http://www.consecol.org/vol7/iss3/art9 Toledo,V . M. et al2004 .E l Potencial Economico de laFlor a utild e loscafetale s de la Sierra Norte de Puebla. Informe Final delproyect o AE019 al Consejo nacional para la Conservation y Uso de la Biodiversidad (CONABIO). Toledo, V.M. and P. Moguel. 2006. Multiple benefits of shade-grown coffee in bioregional conservation strategy Conservation Biology (accepted). Trzcinski, M.K., Fahrig, L. and Merriam, H.G. 1999.Independen t effects of forest cover and fragmentation on the distribution of forest breeding birds. Ecological Applications 9: 586-593. Van der Voort, M. and Greenberg, R. 1999.Wh y Migratory Birds are Crazy for Coffee. Smithsonian Migratory Bird Center. U.S.A. Van Noordwijk, M., Rahayu, S., Hairiah, K., Wulan, Y.C.,Farida , A. And Verbist, B.2002. Carbon stock assessment for a forest-to-coffee conversion landscape in Sumber- Jaya (Lampung, Indonesia): from allometric equations to land use change analysis. Science in China (Series C) 45: 75-86. Vennini, J.P. 1994.Nearti c avian migrants in coffee plantations and forest fragments of southwestern Guatemala, Bird Conserv. Int. 4:209-232 . Wiens,J.A . 1989. Spatial scaling in ecology. Func. Ecol. 3:385-397 . Willis,K.J . and Whittaker, R.J. 2002. Species Diversity - Scale matters. Science295 : 1245-1248. Wunderle, J. and Latta, S. 1996.Avia n abundance in sun and shade coffee plantations and remnant pine forest inth e Cordillera Central, Dominican Republic. OrnitologiaNeotropica l 7: 19-34. 199 Annex Family Species EN S PC FG FD Cracidae Ortalis vetula NA R NA F NA Ardeidae Bubulcus ibis NA R NA C/l NA Cathartidae Coragyps atratus NA R NA C NA Cathartesaura NA R NA c NA Accipitridae Pandion haliaetus NA M NA c NA Accipiter striatus NA M SSP c NA Buleo magnirostris NA R NA c NA Buteo platypterus NA M SSP c NA Columbidae Patagioenas flavirostris NA R NA G/F NA Zenaida asiatica NA R NA G/F NA Columbina inca NA R NA G NA Columbina talpacoti NA R NA G NA Leptotila verreauxi NA R NA G/F NA Geotrygon albifacies NA R E G FR Psittacidae Pionus senilis NA R E F NA Amazona autumnalis NA R NA G/F NA Cuculidae Coccyzuserythropthalmus NA M NA F/l NA Piaya cayana NA R NA 1 NA Crotophaga sulcirostris NA R NA 1 NA Strigidae Megascops guatemalae NA R NA C NA Glaucidium brasilianum NA R NA C NA Ciccaba virgata NA R NA C NA Apodidae Streptoprocne zonaris NA R NA 1 NA Chaeturavauxi NA R NA 1 NA Trochilidae Campylopterus curvipennis NA R NA N/l NA Campylopterushemileucurus NA R NA N NA Anthracothorax prevostii NA R NA N/l NA Hylochoris leucotis NA R NA N NA Amazilia Candida NA R NA N/l NA Amazilia cyanocephala NA R NA N NA 200 Family Species EN S PC FG FD Amazilia tzacatl NA R NA N/l NA Amazilia yucatanensis MX R NA N NA Lampornis amethystinus NA R NA N NA Lampornis clemenciae NA R NA N NA Eugenes fulgens NA R NA N NA Archilochus colubris NA M NA N/l NA Atthis heloisa MX R NA N NA Selasphorus platycercus NA R NA N NA Trogonidae Trogonelegans NA R NA 0 FR Momotidae Momotus momota NA R NA 0 NA Ramphastidae Aulacorhyncus prasinus NA R SSP O NA Picidae Melanerpes formicivorus NA R NA O FR Melanerpes aurifrons NA R NA 1 NA Sphyrapicus varius NA M NA 0 NA Veniliornisfumigatus NA R NA 1 NA Piculus rubiginosus NA R NA 1 NA Dryocopus lineatus NA R NA 1 NA Campephilusguatemalensis NA R SSP 1 NA Dendrocolaptidae Sittasomus griseicapillus NA R NA F/l FR Xiphorhynchus flavigaster NA R NA 1 NA Lepidocolaptes qffinis NA R NA 1 FR Tyrannidae Camptostomaimberbe NA R NA l/F NA Mionectes oleagineus NA R NA F NA Rhynchocyclus brevirostris NA R NA 1 FR Platyrinchus cancrominus NA R SSP 1 NA Mitrephanesphaeocercus NA R NA 1 NA Contopus cooperi NA M NA 1 NA Contopuspertinax NA R NA 1 NA Contopus sordidulus NA M NA 1 NA Contopus virens NA M NA 1 NA Empidonax flaviventris NA M NA 1 NA Empidonax alnorum NA M NA l/F NA 201 Family Species EN S PC FG FD Empidonax traillii NA M NA NA Empidonax hammondii NA M NA NA Empidonax wrightii NA M NA NA Empidonax affinis NA R NA NA Empidonax difficilis NA R NA NA Empidonax occidentals NA R NA NA Sayornis nigricans NA R NA NA Myiarchus tuberculifer NA R NA l/F NA Myiarchus crinitus NA M NA 0 NA Myiarchus tyrannulus NA R NA 0 NA Pitangus sulphuratus NA R NA l/F NA Megarynchus pitangua NA R NA 0 NA Myiozetetes similis NA R NA l/F NA Myiodynastes luteiventris NA R NA 1 NA Tyrannus melancholicus NA R NA 0 NA Tyrannus couchii NA R NA 0 NA Tityra semifasciata NA R NA F NA Vireonidae Vireogriseus NA M NA 0 NA Vireobellii NA M NA 1 NA Vireosolitarius NA M NA O NA Vireo huttoni NA R NA 0 NA Vireogilvus NA M NA 0 NA Vireoleucophrys NA R NA F/l NA Vireo olivaceus NA M NA l/F NA Vireophiladelphicus NA M NA 0 NA Hylophilus ochraceiceps NA R SSP 1 NA Vireolaniusmelitophrys NA R NA 1 NA Cyclarhis gujanensis NA R NA 1 NA Corvidae Cyanocorax yncas NA R NA 0 NA Cyanocoraxmorio NA R NA o NA Hirundinidae Stelgidopteryx serripenis NA R NA 1 NA Hirundo rustica NA R NA 1 NA 202 Family Species EN S PC FG FD Troglodytidae Campylorhynchus zonatus NA R NA NA Catherpesmexicanus NA R NA NA Thryothorus maculipectus NA R NA NA Troglodytesaedon NA M NA NA Henicorhina leucosticta NA R NA NA Henicorhina leucophrys NA R NA FR Regulidae Regulus calendula NA M NA 0 NA Sylviidae Polioptila caerulea NA M NA 1 NA Turdidae Sialia sialis NA R NA 0 NA Myadestes occidentalis NA R SSP F NA Myadestes unicolor NA R E F NA Catharus aurantiirostris NA R NA 0 NA Catharus frantzii NA R E O NA Catharus mexicanus NA R SSP 0 NA Catharus ustulatus NA M NA F NA Hylocichla mustelina NA M NA o NA Turdusgrayi NA R NA l/F NA Turdusassimilis NA R NA l/F NA Mimidae Dumetella carolinensis NA M NA 0 NA Toxostoma longirostre NA R NA F NA Melanotis caerulescens MX R NA F NA Bombycillidae Bombycilla cedrorum NA M NA F/l NA Parulidae Vermivorapinus NA M NA 1 NA Vermivora celata NA M NA 1 NA Vermivoraruficapilla NA M NA 0 NA Parula americana NA M NA 0 NA Parula pitiayumi NA R NA 1 NA Dendroica magnolia NA M NA 1 NA Dendroica coronata NA M NA 1 NA Dendroica virens NA M NA 1 NA Dendroica townsendi NA M NA 1 NA Dendroica occidentalis NA M NA 1 NA 203 Family Species EN PC FG FD Mniotilla varia NA M NA NA Setophaga ruticilla NA M NA NA Helmitheros vermivous NA M NA NA Seiurus aurocapilla NA M NA NA Seiurusmotacilla NA M NA NA Oporornisformosus NA M NA NA Geothlypistrichas NA M NA NA Geothlypispoliocephala NA R NA O NA Wilsonia citrina NA M NA NA Wilsoniapusilla NA M NA NA Wilsoniacanadensis NA M NA NA Myioborus miniatus NA R NA NA Euthlypis lachrymosa NA R NA NA Basileuterus culicivorus NA R NA NA Basileuterus rufifrons NA R NA NA Basileuterus belli NA R NA NA Icteria virens NA M NA 0 NA Thraupidae Chlorospingus ophthalmicus NA R NA 0 NA Piranga rubra NA M NA O NA Piranga ludoviciana NA M NA O NA Piranga leucoptera NA R NA 0 NA Thraupisabbas NA R NA F NA Cyanerpes cyaneus NA R NA 0 NA Emberizidae Volatiniajacarina NA R NA G/l NA Sporophila americana NA R NA G NA Sporophila torqueola NA R NA G NA Sporophila minuta NA R NA G NA Tiaris olivaceus NA R NA 0 NA Atlapetes albinucha NA R NA 1 NA Buarremon brunneinucha NA R NA 1 FR Arremonops rufivirgatus NA R NA G NA Aimophila rufescens NA R NA G NA 204 Family Species EN S PC FG FD Spizella passerina NA R NA G/l NA Melospiza lincolnii NA M NA G/l NA Cardinalidae Saltator atriceps NA R NA 0 NA Pheucticus ludovicianus NA M NA 0 NA Cyanocompsaparellina NA R NA 0 NA Passerina caerulea NA R ND G NA Passerina cyanea NA M NA G NA Passerina ciris NA M NA G NA Icteridae Dives dives NA R NA 0 NA Quiscalus mexicanus NA R NA 0 NA Molothrus aeneus NA R NA G/l NA Icterus spurius NA M NA 0 NA Icterus cucullatus NA M NA F/l NA Icterus gularis NA R NA l/N NA Icterus graduacauda MX R NA l/N NA Icterus galbula NA M NA O NA Amblycercus holosericeus NA R NA 1 NA Psarocolius montezuma NA R SSP F NA Fringillidae Euphonia hirundinacea NA R NA G/F NA Euphonia elegantissima NA R NA F NA Carduelispinus NA R NA G NA Carduelispsaltria NA R NA G NA Coccothraustes abeillei NA R NA G NA Annex 1. Totallist of observed,heard and/or captured birds in taxonomic order inthe study area (according toAOU1998). Abbreviations: EN, Endemic; PC, Protection Category;FG, Foraging Guild; FD, Forest Dependent; MX, endemic toMexico; R, resident;M, migratory;E, endangered; SSP,subject tospecial protection; F.frugivorous; I, insectivorous; G, granivorous, N, nectarivorous; O, omnivorous; C,carnivorous; FR, forest dependent (species that breedexclusively in theforest); NA, not applicable. 205 m» & ;r--'. Summary Resumen Samenvatting Summary As the current accelerated and increasing loss ofbiologica l diversity have become apparent land managers and ecologists have sought to identify significant habitats toth e preservation ofbiodiversity . A critical component ofbiodiversit y protection isth e understanding of the ecological forces shaping the species diversity patterns. The aim of this study ist o gain insight inth e local and regional factors ultimately controlling species persistence and coexistence. The conceptual background ofthi s study istha t of a diversified multiple-use landscape matrix, that isuse d and managed andwher e "natural areas"ca n be embedded. Neotropical bird species are currently under threat as their breeding grounds suffer from degradation and loss because of intensification of land use. The fieldwork ofthi s thesis was conducted inth e northeastern mountain range of Puebla, Mexico. This region represents an important area for the conservation ofresiden t and migrant birds, as it is located ina strategic position at theNearti c and Neotropical biogeographic boundary. It also forms part of the migratory route for Nearctic-Neotropical birds. Moreover, despite the loss of primary forests in the region, one of the main land uses in the study area is traditional shade coffee plantations which remain as an important forested habitat for birds. An important issue in conservation biology is the monitoring of community trends to provide reliable information of species diversity and their status, for fast and efficient identification of conservation priorities. In chapter 2, we analyse the use of a double- observer point count approach and mist netting for assessing bird species richness during migration. We assess the relative biases, costs and efficiency of both techniques to aim optimisation in the design of large-scale monitoring. We found that the double-observer point count technique was the most effective in the total species richness completeness and presented lower total effort in comparison to mist netting. The performance ofpoin t counts is higher than mist netting in the detection ofne w bird species in the research area, even after a large sampling effort. However, mist netting significantly detected a higher proportion of understory species in comparison to point counts, though we found opposite results for migrant species. Finally, the cost-efficiency analysis showed that the modified 209 double-observer point counts required lesstota l effort thus decreasing total monetary costs compared to mist netting. One ofth e main problems in conservation in Latin America isth e accelerated deforestation and conversion to monocultures and grazing lands that has direct effects on Neotropical avian communities (i.e.residen t and migrants) leading to amajo r loss ofhabitat , and landscape fragmentation. In chapter 3,w e analyse how fragmentation and habitat loss inth e landscape influences the bird species richness patterns.W e examine the relative individual and combined influence of these two factors on species richness in an avian metacommunity. Moreover, we compare the difference in explanatory power of individual and combined influence of both factors. The response of species richness to habitat fragmentation shows aunimoda l response at landscape level, and a negative response to habitat loss. The combined influence of fragmentation and habitat loss did not offer a better approximation of species richness response. This suggests that there isn o interaction between the effects of fragmentation and habitat loss.Assessmen t of the effects of habitat fragmentation and loss under the current situation of growing human perturbation in natural habitat is fundamental in conservation and landscape management. An important ecological force structuring ecological communities is interspecific competition. Body mass isa n easily determined characteristic of animals that probably influences competition strength. In chapter 4, our objective is to examine the effect of body size (mass) on competitive interactions between competing pairs ofbir d species. Our results indicate that there is a significant negative relationship between bird body mass ratio and the competition strength i.e.; the larger the body mass ratio,th e lower the competition strength thereby suggesting that high variation in body sizes amongst sympatric species may promote coexistence in communities. Moreover species that have a greater overlap in resource use tend to exhibit stronger competition than species that overlap less in their resource use. 210 In chapter 5,th e influence of spatially explicit bio-physical variables at multiple scales on a bird community is analysed. We argue that biological communities are organised at multiple functional spatial scales and interactions between these scales determine both local and regional patterns of species richness. We use a multiple scale approach with plot, patch and landscape level variables using abundance and presence-absence data. Our results demonstrate that landscape variables explain most of the variation in bird species in both abundance and presence-absence analyses in all explanatory sets.Interestingly , results demonstrate that variation in community structure was described best at family-level than at genera- or species-level. Our results show that shade coffee plantations is one of the main land covers thatpositivel y influence the species richness, thusprovidin g habitat for neo tropical migrants and forest-dependent birds (e.g.; in this study some endemic and protected species). Thus, selecting the appropriate scale(s) of management in conservation strategies isessentia l in conservation of bird communities. The use of remotely sensed data has great potential to aid in explaining species diversity and community assemblage patterns at multiple scales. Besides, it can help to optimise sampling strategies or to allow testing of hypotheses regarding the spatial correspondence of species diversity patterns among different taxonomic groups. In chapter 6,w e review how remote sensing has been used to assess terrestrial faunal diversity, with emphasis on proxies and methodologies, while exploring prospective challenges for the conservation and sustainable use of biodiversity. We grouped and discussed papers dealing with the faunal taxa mammals, birds,reptiles , amphibians, and invertebrates into five classes of surrogates of animal diversity: 1.habita t suitability, 2.photosyntheti c productivity, 3. multi-temporal patterns,4 . structural properties ofhabitat , and 5. forage quality. It is concluded that the most promising approach for the assessment, monitoring, prediction, and conservation of faunal diversity appears to beth e synergy ofremot e sensing products and auxiliary data with ecological biodiversity models, and a subsequent validation ofth e results using traditional observation techniques. 211 In chapter 7,1presen t the conservation implications of shade coffee plantations as a refuge for Neotropical birds. The scenarios of loss and conversion of shade coffee plantations for biodiversity conservation and carbon sequestration emphasise on the ecological services that this agro-ecosystem provides. Thereductio n of shade coffee plantations in the studied region will have a deleterious effect onth e species richness of e.g., resident birds with an estimated decline of 0.006 in species richness for every hectare the coffee area is reduced. These results could have significant consequences for conservation strategies as the amount oftraditiona l shade coffee plantations has apositiv e linear relationship with the species richness of resident birds (p<0.001) . Moreover, in our research area the calculated carbon stock ofth e total area covered by shade coffee plantations is 257,7891C, which represents 1,288,943US D in carbon credits. Thus, with areporte d current conversion rate of 0.4 % the results in carbon loss will be devastating injus t one year. Although the conversion rate for the research area is unknown we approximate an almost negligible rate of conversion compared to the previously reported. The potential for the coffee producers as suppliers of carbon sequestration is calculated in approximately 105USD/ha , which represents a considerable income compared to approximately an average of 59USD/h a per coffee harvest. In addition, the value of shade coffee plantations as a component of the anthropogenic matrix is stressed in this study. Traditional shade coffee plantations are cultivated mainly by small-scale community-based growers,th e majority of them belong to some indigenous group. The indigenous form of utilisation of thistropica l agroecosystems denotes a multiple use strategy, which can be called adaptative management. This indigenous adaptative management has obvious socio-economic aswel l as ecological benefits. The ecological advantages of shade coffee agroecosystems are clear: a) high biodiversity maintenance, b) regulation of carbon cycle, c) soil protection, d)regulatio n of hydrological cycle and e) preservation of forest cover. The economic benefits derived from this tropical agroecosystems are sources of goods, services and energy for household subsistence and products for local, regional and international markets. To increase the effectiveness of conservation management, the value of suitable land for multiple uses, as a component of 212 the anthropogenic matrix should be considered. For example, inth e case oftropica l bird species, such linkages could bemaintaine d by using the man-modified landscapes, e.g.; traditional shade coffee plantations asthes e areas also harbour an ecological value as habitat 213 Resumen Debido a la acelerada y creciente perdida de la diversidad biologica (subsecuentemente referida como biodiversidad) los manejadores de recursos naturales y ecologos han buscado identificar los habitats relevantes para lapreservatio n de la biodiversidad. Un componente critico para laprotectio n de labiodiversida d es el entendimiento de las fuerzas que configuran lospatrone s de ladiversida d de especies. La meta de este estudio es el de comprender los factores locales y regionales que finalmente controlan la persistencia y coexistencia de las especies. El concepto sobre el cual se fundamenta este estudio es hacia el uso y manejo deun a matriz diversificada del paisaje con multiples usos en donde las "areas naturales" se encuentran insertas. Las avesNeotropicale s se encuentran actualmente amenazadas ya que sus areas de reproduction sufren degradation yperdida . El trabajo de camporealizad o en esta tesis fue conducido en la Sierra Norte de Puebla, Mexico. Esta region representa un area importante para laconservatio n de aves residentes y migratorias, ya que se encuentra localizada en una zona estrategica biogeografica en el limite entre la region Neoartica y la Neotropical. Ademas, forma parte del corredor migratorio para las avesNeoarticas-Neotropicales . A pesar de la alta tasa de deforestation en la region, uno de los usos de sueloprincipale s en el area de estudio esplantacione s tradicionales de cafe bajo sombra los cuales permanecen como importantes areas forestadas que representan habitat adecuado para la avifauna. Una cuestion importante en la conservacion es el monitoreo de las tendencias en las comunidades de especies para obtener information confiable de la biodiversidad y su estatus.E n el capitulo 2, analizamos el uso del enfoque de puntos de conteo con doble- observador y el uso de redes de niebla para evaluar lariquez a de especies durante la epoca de migration. Evaluamos las biases y los costos relativos, y la eficiencia de ambas tecnicas para conseguir la optimization en el diseno de monitoreos agra n escala. Nuestros resultados muestran que latecnic a depunto s de conteo con doble-observador tiene un mejor desempefio en la detection del numero total de especies y requiere un esfuerzo de 215 muestreo menor comparado a las redes de niebla. El desempeno de lospunto s de conteo es mejor que las redes de niebla en la detection de nuevas especies, aun despues deu n esfuerzo de muestreo relativamente largo. Sin embargo, las redes deniebl a detectaron significativamente una proportion mayor de especies del sotobosque en comparacion a los puntos de conteo,n o obstante encontramos resultados contrarios enrelatio n a las especies migratorias. Finalmente, el analisis de costo-beneficio mostro que lospunto s de conteo con doble-observador requieren de un esfuerzo de muestreo total por lotant o los costos monetarios son menores en comparacion a las redes de niebla. Uno de losproblema s principales en America Latina es la acelerada deforestation y conversion a monocultivos ypotrero s lo cual tiene efectos directos en las comunidades de avesNeotropicale s (i.e.;ave s residentes y migratorias), lo que conlleva a la fragmentation y perdida substantial de habitat. En el capitulo 3,analizamo s como la fragmentation y perdida de habitat inffuencia los patrones de riqueza de especies. Examinamos la influencia relativa individual y combinada de ambos factores en lariquez a de especies en una metacomunidad de aves. Ademas, comparamos la diferencia del poder explicativo del efecto individual y combinado. Larespuest a en lariquez a de especies a la fragmentation del habitat muestra una respuesta unimodal a nivel del paisaje, y una respuesta lineal negativa a laperdid a de habitat. El efecto combinado de ambos factores (i.e.; fragmentation y perdida de habitat) no ofrece una mejor aproximacion de larespuest a en lariquez a de especies. Esto sugiere que no hay interaction entre los efectos de fragmentation y perdida de habitat. La evaluation de los efectos de fragmentation y perdida de habitat bajo la situation actual de perturbation humana en loshabitat s es fundamental en el manejo y conservation de especies anive l del paisaje. Una fuerza ecologica importante que estructura las comunidades de especies es la competencia interespecifica. La masa corporal es una caracteristica animal facilmente determinable que probablemente influencia la intensidad de la competencia. En el capitulo 4, nuestro objetivo es examinar el efecto de la masa corporal en las interacciones de competencia de pares de especies. Los resultados indican que existe una relation lineal 216 negativa entre latas a de masa corporal y la intensidad de competencia, i.e.; entre mas grande latas a de masa corporal, menor la intensidad de competencia por lo que esto sugiere que la mayor variacion en la el tamano de especies entre especies simpatetricas puede promover la coexistencia en las comunidades de especies.Ademas , las especies que tienen un sobrelapamiento en el uso de recursos tienden a exhibir un grado de competencia mayor que las especies con menor sobrelapamiento en elus o de sus recursos. En el capitulo 5, analizamos la influencia de variables espaciales bio-fisicas en una comunidad de aves,baj o un contexto multi-escalas.Nosotro s argumentamos que las comunidades de especies estan organizadas en multiples escalas espaciales funcionales y las interacciones entre estas escalas determinan lospatrone s locales y regionales de la riqueza de especies. Usamos un enfoque de escalas multiples con variables anive l de parcela, parche y paisaje utilizando datos de abundancia y presencia-ausencia de especies. Los resultados demuestran que lasvariables al nivel delpaisaj e explican lamayo r parte de la variacion en los datos de abundancia y presencia-ausencia de especies de aves en todos los grupos de variables independientes. Interesantemente, lavariacio n en la estructura de la comunidad es descrita mejor a nivel de familia taxonomica que anive l de genero o especie. Tambien, las plantaciones de cafe esun o de las coberturas de suelo que muestran una relation lineal positiva con lariquez a de especies,po r lo que proveen habitat adecuado para aves migratorias y dependientes delbosqu e (e.g.; en este estudio algunas especies endemicas y especies protegidas). Por lo que el seleccionar la escala(s) espacial apropiada en las estrategias de manejo y conservacion puede tener implicaciones importantespar a la conservacion de las comunidades de aves en la region de estudio. El uso de datos obtenidos atrave s de sensores remotos tiene un gran potencial auxiliando en explicar la diversidad de especies y lospatrone s de la estructura en las comunidades de especies en multiples escalas.Ademas ,pued e auxiliar en la optimization de las estrategias de muestreo opermitiend o laverificatio n de hipotesis concernientes a la correspondencia espacial de lospatrone s de ladiversida d de especies entre diferentes grupos taxonomicos. En el capitulo 6,realizamo s una revision de como los sensores remotos han sido usados 217 para evaluar ladiversida d de fauna terrestre,haciend o enfasis en las variable substitutas utilizadas para explicar la distribution y diversidad de especies, al igual que las metodologias usadas. Tambien damos una perspectiva de losdesafio s del uso de datos obtenidos atrave s de sensores remotospar a la conservation y uso sustentable de la biodiversidad. Agrupamos y discutimos articulos cientificos acerca de diferentes grupos taxonomicos, i.e.; mamiferos, aves,reptiles , anfibios, e invertebrados en cinco grupos de variables sustitutas utilizadas para explicar la distribution y diversidad de especies: 1. adecuacion de habitat, 2.productivida d primaria, 3.patrone s multi-temporales,4 . propiedades estructurales delhabitat , y 5. calidad de forraje. Concluimos que el enfoque masprometedo rpar a laevaluation , monitoreo, prediction, y conservation de la diversidad de especies de fauna terrestre resulta ser la sinergia de productos derivados de sensores remotos y datos auxiliares con modelos ecologicos de labiodiversidad , y la subsiguiente validation de losresultado s utilizando tecnicas de observation tradicionales. En el capitulo 7,present o las implicaciones de conservation de los cafetales tradicionales bajo sombra comorefugio s para las avesNeotropicales . Los escenarios de perdida y conversion de los cafetales tradicionales bajo sombra mostrados en este estudio, enfatizan sobre los servicios ecologicos que estos agroecosistemas proveen. La reduction de los cafetales bajo sombra en laregio n de estudio tendria un efecto perjudicial en la riqueza de especies de,po r ejemplo, aves residentes, con una reduction en la riqueza de especies de 0.006po r cada hectarea perdida de cafe bajo sombra. Estos resultados pueden tener consecuencias significativas para las estrategias de conservation ya que la cantidad de area de cafetales bajo sombra tiene una relation lineal positiva con lariquez a de especies de aves residentes (pO.OOl). Ademas, en nuestra area de estudio el abastecimiento calculado de carbono producido por el area total cubierta por cafetales bajo sombra es de 257,7891C lo querepresent a 1,288,94 3 USD en creditos de carbono. Por lo que si consideramos la tasa de conversion reportada para los cafetales bajo sombra en Mexico de 0.4% los resultados en laperdid a de carbono serian devastadores en solamente un afio. Sin embargo, latas a de conversion en laregio n de estudio es desconocida y estimamos que comparada con la tasa reportada relativamente pequena. El potencial monetario para losproductore s de 218 cafe como abastecedores de secuestracion de carbono esta calculado en aproximadamente en 105USD/ha , lo cual representa un substancial comparado con promedio aproximado de 59 USD/ha por cosecha de cafe. Mas aun, elvalo r de las plantaciones de cafe bajo sombra como un componente de la matrizantropogenica se enfatiza en este estudio. Los cafetales tradicionales bajo sombra son cultivados mayoritariamente por pequefios productores ejidatarios, la mayoria de ellosperteneciente s a algun grupo indigena. Elmod o de utilization indigena en Mexico en estos agroecosystemas tropicales denota una estrategia deus o multiple (EUM), la cual ha sido llamada manejo adaptativo. La EUM tiene beneficios socio-economicos obvios como tambien ecologicos. Las ventajas ecologicas de estos agroecosystemas es claro: a) mantenimiento denivele s altos de biodiversidad, b) regulacion de ciclos de carbono, c)protectio n del suelo, d)regulacio n de ciclos hidrologicos, y e)preservatio n de la cobertura forestada. Los beneficios economicos derivados de este agroecosystema son como fuente debiene s y servicios,y energia para la subsistencia de losproductores , y como fuente de productos para mercados locales, regionales e internacionales. Para incrementar la efectividad del manejo y conservation de especies, el valor deuso s de suelo adecuados de usos multiples como un componente de la matriz antropogenica debe ser considerada. Por ejemplo, en el caso de las especies de aves Neotropicales,tale s vinculos pueden ser mantenidos usando los paisajes modificados, e.g.; los cafetales tradicionales bajo sombra ya que estas areas tienen un altovalo r ecologico como habitat adecuado. 219 Samenvatting Sinds de huidige afname van biologische diversiteit duidelijk werd, hebben natuurbeheerders en ecologen getracht die habitats te identificeren, die belangrijk zijn voor het behoud van diebiodiversiteit te behouden. Ombiodiversiteit te behouden is een goed begrip van de ecologische processen diepatrone n in soortsdiversiteit veroorzaken onontbeerlijk. Het doel van deze studie is inzicht verwerven in de ultimate lokale en regionale factoren die het voorkomen en de coexistentie van soorten bepalen. De conceptuele achtergrond van de studie is het interpreteren van een landschap als een gevarieerde multifiinctionele landschapsmatrix, waarin natuurlijke gebieden een plaats wordt gegeven. Vogels in de neotropische zone staan onder bedreiging: hun broedgebieden raken gedegradeerd, of raken ongeschikt. De data voor deze studie werd verzameld in het noordoostelijke berggebied van Puebla, Mexico. Deze regio isee n belangrijk gebied in het beheer van zowel stand- als trekvogels, aangezien het op de grens ligt van deNeartisch e en deNeotropisch e biogeografische zones. Daarnaast ligt het op een belangrijke trekroute van Neartische enNeotropisch e migrerende vogels. Ondanks het verlies van primair bos in deze regio, biedt het gebied nog steeds een belangrijk boshabitat, in de vorm van koffieplantages, waar de traditionele verbouwing van bosrijk schaduwplantage nog wordt toegepast. Voor een passend en efficient natuurbeheer ismonitorin g van soorten essentieel: alleen zo verkrijgen beleidsmakers betrouwbare informatie over soort-diversiteit en status van soorten. In hoofdsruk 2 van dit proefschrift evalueren we twee monitorings-methodes: punt- tellingen mettwe ewaarnemer se n inventarisaties met behulpva n mistnetten. We vergelijken detwe e methodes wat betreft gevonden resultaten, maar ook wat betreft kosten en. Met deze kennis kunnen grote-schaal monitoring geoptimaliseerd worden. Punt- tellingen bleek het meest effectief: deze gaf debest e schatting van de totale soortenrijkdom en kostte minder inspanning. Punt-tellingen werkten ook beter in het vaststellen van nieuwe 221 soorten dan de mist-net methode. Met mistnetten werden relatief meer soorten gevonden die in struwelen leven,maa r depunt-tellinge n gaf meer migrerende soorten. Kost-efficiency analyses toonde aan dat monitoring met punt-tellingen minder inspanning en dus minder geld kosten. Een van debelangrijkst e problemen in natuurbeheer in Latijns Amerika is de toenemende ontbossing. Bossen worden omgevormd tot monocultuur-plantages ofweide . Dit proces lijdt tot verlies of fragmentatie van habitats, en heeft zo een sterke invloed op zowel migrerende alsniet-migrererend e neotropische avifauna. In hoofdstuk 3analysere n we de invloed van fragmentatie en habitatverlies oppatrone n van soortdiversiteit van deze avifauna. We bepalen de individuele en gecombineerde invloed van deze twee factoren op soortsrijkdom in de metagemeenschap van vogels. Soortenrijkdom reageert unimodaal op fragmentatie van habitats of landschapsniveau, en negatief op habitat-verlies. Het gecombineerde effect van deze twee factoren verklaarde het patroon van soortenrijkdom niet beter. Dit suggereerd dat ergee n interactie istussen deeffecte n van fragmentatie en verlies van habitat. Met oog op de groeiende verstorende invloed van de mens in natuurlijke habitats is en beoordeling van de effecten van fragmentatie en verlies van habitat is een van groot belang bij landschapsbeheer. Interspecifieke concurrentie isee n belangrijke onderdeel van het proces van vorming van ecologische gemeenschappen. Lichaamsgrootte isee n eenvoudig te bepalen eigenschap en is in potentie een belangrijke bepalende factor in concurrentie tussen soorten.. In hoofdstuk vier bestuderen we het effect van lichaamsgrootte (massa) op deze interactie tussen vogelsoorten. Onze resultaten tonen dat er een significante negatieve relatie bestaat tussen lichaamsgewicht van soorten en de mate waarin deze soorten concurreren: hoe hoger de lichaamsgrootte-ratio hoe lager de competitie-kracht. Dit suggereert dat de grote variatie in lichaamsgrootte in sympatrische soorten de coexistentie in gemeenschappen mogelijk maakt.Bovendie n tonen soorten met een grotere overlap ingebrui k van resources een sterkere concurrentie dan soorten die hierin minder overlap tonen. 222 In hoofdstuk 5 analyseren we de invloed van ruimtelijk expliciete biologische en fysieke variabelen opvogel-gemeenschappen . We doen dit op verschillende ruimtelijke schalen, omdat we verwachten datbiologisch e gemeenschappen op meerdere functionele ruimtelijke schalen georganiseerd zijn en dat interacties tussen deze schalen de locale en regionale patronen in soortsrijkdom bepalen. We doen dit door de abundantie en aan- afwezigheids- van de vogelsoorten te analyseren aan de hand van variabelen op het niveau van plot, patch of landschap. Onze resultaten tonen dat zowel abundantie als aan- afwezigheid van soorten het beste verklaard worden door variabelen die spelen op landschaps-schaal. Interessant is dat onze resultaten uitwijzen dat variatie in de structuur van de gemeenschappen het best beschreven wordt op het taxonomische niveau van familie, beter dan op genus of soorts-niveau. Traditionele koffieplantages vormen een van de landschaps-types die soortenrijkdom positief beinvloeden, en vormt zo een habitat voor zowel neo-tropische migranten als bosbewondende vogels, waaronder een aantal endemische, beschermde soorten. Het selecteren van dejuist e geografische schaal bij opstellen en implementeren van beschermingsplannen voor vogel-gemeenschappen is uiterst belangrijk. Het gebruik van remote sensing voegt opent nieuwe mogelijkheden in het onderzoek aan patronen in soortenrijkdom en structuur van gemeenschappen opverschillend e schalen. Daarnaast kan het een goed gereedschap zijn bij het optimaliseren van bemonstering van soortsdiversiteit, ofteste n van hypotheses over relaties tussen soortenrijkdom in verschillende taxonomische groepen. In hoofdstuk 6bespreke n we hoe remote sensing wordt gebruikt om faunistische diversiteit te bepalen, met een nadruk op methodologie. We bespreken mogelijke uitdagingen op het gebied van beheer en duurzaam gebruik van biodiversiteit. We delen de studies van zoogdieren, vogels reptielen amfibieen en invertebraten op in 5 klassen van surrogaten voor daadwerkelijke-diversiteit: 1) geschiktheid van habitat, 2) fotosynthetische productiviteit, 3)tijdsreeks-patronen . 4). structurele eingeschappen van het habitat en 5)voedselkwaliteit . We concluderen dat de meest kansrijke aanpak van bepalen, monitoren, voorspellen enbehere n van diversiteit in fauna gelegen is in de synergie van remote sensing en ecologische biodiversiteits-modellen, 223 gevolgd door een validatie van de resultaten met behulp van de meer traditionele veld- methodiek. In hoofdstuk 7bespree k ik de implicaties die traditionele koffieplantages met schaduw- bouw, in hun functie alsrefugiu m voor neotropische vogels,hebbe n op natuurbeheer. De scenarios van versies of omzetting van deze koffieplantages voor het beheer van biodiversiteit enkoolstof-fluxe n benadrukken de ecologische rol van dit agro-ecosysteem. De afname van koffieplantages inhe t studiegebied heeft een negatief effect op soortenrijkdom van standvogels, met een geschatte afname in soortenrijkdom van 0.0006 voor elkehectar e plantage die verdwijnt. Deze resultaten kunnen behoorlijke consequenties hebben voor beheer-strategieen aangezien het oppervlakte aan koffieplantages een positieve lineaire relatie heeft met soorten-rijkdom van standvogels (P < 0.001). Bovendien isvoo r ons studiegebied te berekenen dat in deze koffieplantages 257,7891C aan koolstof isvastgelegd, , wat correspondeert met 1,288,943US D in carbon credits. De huidige omzetting van plantages naar niet-bebosde landschappen van 0.4% perjaa r voor heel Mexico, zou voor het studiegebied leiden tot een aanzienlijk verlies van gefixeerd koolstof aanzienlijk. Hoewel de omzettingssnelheid in het studiegebied onbekend is,voorspelle n we een bijna verwaarloosbare omzettingssnelheid. Koffieproducenten zouden een koolstoffixatie kunnen leveren van ongeveer 105US D ha"1,wa t een hoger inkomen zou geven dan de gemiddelde winst bij verbouwens van koffie (59 USD ha"1). Daarnaast hebben deplantage s waarde als antropogeen element van de landschaps-matrix: traditionele schaduwplantages worden beheerd door kleine kleine boeren uit lokale kleine gemeenschappen, waarvan ervele n behoren tot een inheemse bevolkingsgroep. De inheemse wijze van gebruik van dit tropische agroecosysteem volgt een multiple use strategie, ook wel adaptive management genoemd. Zulkeadaptive management strategieen kennen zowel socio-economische als ecologische voordelen. De ecologische voordelen van traditionele koffieplantages zijn duidelijk: a)z e bieden ruimte aan een hoge biodiversiteit. b) ze leggen koolstof vast. c).z e beschermen tegen bodem erosie. d) ze gaan uitdroging tegen. e)z e maken behoud van bos mogelijk. Vanuit een economisch perspectief vormen de koffieplantages een bron van goederen, service en energie voor lokale bewoners en van 224 producten voor lokale,regional e en intemationale handel. Omeffectivitei t van het beheer te vergroten, moet waarde van dit soort landschappen in met meerdere functies worden onderkent, en moet een gebied worden gezien als onderdeel van de antropogene matrix. In het gevalva n tropische vogels kunnen zulkeverbindinge n kunnen worden beheerd door gebruik van die onderdelen van de landschapsmatrix die gebruikt worden door de mens, zoals traditionele schaduw-bebouwede koffieplantages, aangezien deze ook functioneren als habitat. 225 ##•*%'., ••4 »• v ^ 1 Acknowledgments Curriculum vitae List of publications Affiliations of co-authors PE&RC PhD Education Statement Form Acknowledgments I am grateful to allpeopl e mentioned here for their intellectual, moral and material support. I express my gratefulness to Fred de Boer for his guidance, supervision and faith in my work. I also feel thankful for all the support given by Antoine Cleef during the long process of my PhD.M y gratitude toVicto r Toledo for introduced me to the coffee lands and giving me support when needed. My thanks to Andrew Skidmore for his supervision. Without Samuel Lopez de Aquino that Iwouldn' t have been able to carry out my fieldwork, and Iappreciat e all the knowledge and friendship he offered me. I especially thank all the people ofth e Cooperative Tosepan Titataniske for allth e help, support and honest interest in my study that made my fieldwork possible and that ended in a creative and productive process beyond science. I express my gratitude to the people in Cuetzalan region for making me feel athome , especially Dona Guille. My appreciation to Patricia Moguel for initiating the topic of mythesis . Iappreciat e the help given by the excellent birdwatchers, Garry Bakker and Reinoud Vermoolen,durin g their visit to my research area. I am grateful to Javier Lopez Caloca and Mario Gomes Soberon for providimg the digital geographic material.Nesto r Matamoros, Sofia Morales Argiiello, Yolanda Aguirre Montiel and Sandra Jimenes Peregrina assisted in processing the digital material. Lucia Almeida, Victor Avila-Akerberg, Veronica Aguilar, Tanya Gonzalez, Jose Hernandez, Maarten Blaauw and Ernesto Anaya contributed with the vegetation survey. Nelly Garcia Diego and Beatriz Ludlow performed the taxonomic identification ofplan t material. Frank van Langevelde, Gerard Reinink and Boudewijn van Leeuwen helped process the geographic information. My gratitude goes aswel l to Idea Wild for the field material provided for my research. Iappreciat e the help and camaraderie given by all the staff and PhD students at the Resource Ecology Group. Thanks to the family Drent for all their support. 229 I am indebted to my parents, Fernando and Chuy, for providing me with love and the moral and material environment that allowed met oreac h this stage.Thank st o my friends, Bibiana, Rodrigo, Michiel, Manuel, Sergio,Emiliano , Adolfo, Lilia, Maarten, Max. Thanks to all my past andpresen t housemates:Jasja , Ans, Jasper, Elise, Vena, Paul, Karin, Griet, Arthur, Sander, Maartje. Ia m thankful to Annelies and Ton for being honest, natural and loving. Tijl, Ia m so grateful for your love, sharp mind, humour, patience, understanding, critics, fun, and for being there even in my worse moments. And finally, Ihav e to mention Yohuak, my faithful dog that no matter what, he drags me out to the sun orrain , giving my lungs and mind a moment of fresh air. Euridice Wageningen, March 2006 230 Curriculum vitae Euridice Leyequien Abarca was born on September 21st 1973 in Mexico City, Mexico. In 1997 she obtained her BSc. Degree in Biology from the Universidad Autonoma Metropolitana Unidad Xochimilco (UAM-X), Mexico City, Mexico. During her studies at UAM-X she conducted a land use planning study based on forestry goals in the municipality Ecatzingo de Hidalgo, Estado de Mexico, Mexico. After graduating, she worked as assistant professor in the same University in the Laboratory of Geographical Information Systems. At the end of the year 1998 she started her MSc. Degree in Environmental System Analysis and Monitoring from the International Institute for Aerospace Survey and Earth Sciences (ITC), Enschede, the Netherlands. Her MSc. thesis was conducted in the Hurungwe and Nyaminyami rural districts, Zimbabwe and it dealt with the spatial and temporal assessment of grazing and water resources for livestock in communal lands. She obtained her MSc. Degree in 2000. After graduating, she returned to Mexico and started to work inth e project 'Etno-ecological Atlas for Mexico and MidAmerica" in the Laboratory of Etnoecology, Institute of Ecology (now Research Centre in Ecosystems), Universidad Nacional Autonoma de Mexico (UNAM). In 2002 she started her PhD in the Resource Ecology Group from Wageningen University, the Netherlands. Her research project entailed the study ofth e ecological forces shaping the bird species diversity, and was performed in the north eastern mountain range ofPuebla , Mexico in a coffee region. Her research allowed her to collaborate with an indigenous Cooperation Tosepan Titataniske which groups coffee farmers with organic and/or traditional coffee production. The results of this thesis contribute toth e conservation and management plans ofth e aforementioned coffee region that aims to a diversified multiple-use in the landscape matrix, where "natural areas"ca n be embedded. 231 List of Publications Published or accepted Toledo, V.M.; Alarcon-Chaires, P.; Moguel, P.; Olivo, M.; Cabrera, A.; Leyequien, E. y Rodriguez-Aldabe, A. 2001.E l Atlas Etnoecologico De Mexico y Centroamerica: Fundamentos, Metodos y Resultados. 6 (8): 7-41. Toledo, V.M.; Alarcon-Chaires, P; Moguel, P.; Olivo, M; Cabrera, A.; Leyequien, E. y Rodriguez-Aldabe, A. Biodiversidad y Pueblos Indigenas En Mexico y Centro America. 2002. Biodiversitas. 7 (43):2-8 . Chavez-Cortes, J.M., Chavez-Cortes, M.M., Binnquist-Cervantes, G.S., Roldan-Aragon, I., Leyequien, E., Romano-Delon, G. 2003. Strategic Planning Of The Iztaccihuatl - Popocatepetl Area, In: Heil, G.W., Bobbink, R and Trigo Boix Nuri (Eds.) Ecology and Man in Mexico's Central Volcanoes Area. Geobotany 29, Kluwer, Dordrecht Toledo,V.M. , Moguel, P., Duran, L.,Albores , Ma. L., Leyequien, E., Ayon, A., Rodriguez- Aldabe, A. and Alarcon-Chaires, P. (Accepted in Etnobiologia 2004). Etnobiologia para la Resistencia Indigena: La Sierra Norte de Puebla. Leyequien, E., Verrelst, J., Slot, M, Heitkonig, I., Schaepman, G. and Skidmore, A.K. (Accepted in International Journal of Applied Earth Observation and Geoinformation, Special Issue). Capturing the Fugitive: Applying Remote Sensing to Terrestrial Animal Distribution and Diversity. A Review 233 Submitted Leyequien, E.,D e Boer, W.F. and Skidmore,A.K . Linking Species-Environment Relationships and Multiple Spatial Scales in Community Ecology. Leyequien, E., De Boer, W.F. and Toledo, V.M. Fragmentation and Habitat Loss: Species Richness Responses in an Avian Metacommunity. Leyequien, E., De Boer, F. and Cleef, A.M. Influence of Body Size on Coexistence of Bird Species. Leyequien, E., De Boer, F., Bakker, G., Vermoolen, R. and Lopez de Aquino, S. (Submitted to Biodiversity and Conservation). Monitoring Neotropical Birds: Efficiency of a Modified Double-Observer Point Count Approach versus Mist- Netting. In preparation Leyequien, E. and Lopez De Aquino, S.Avifaun a Diversity In The Coffee Agro-Ecological Zone in the Cuetzalan Region, Puebla, Mexico. Leyequien, E., Toledo, V.M. and Lopez de Aquino, S. Etno-Ornithology in the North Eastern Mountain Range of Cuetzalan Region, Puebla, Mexico. Leyequien, E. and Lopez de Aquino, S. The North eastern mountain range of Puebla, Mexico: apriorit y conservation area for Neotropical birds. 234 Affiliation of co-authors Fred de Boer Resource Ecology Group, Department of Environmental Sciences, Wageningen University, Bornsesteeg 69,670 8 PD Wageningen, the Netherlands Antoine M. Cleef Resource Ecology Group, Department of Environmental Sciences,Wageninge n University, Bornsesteeg 69,670 8 PD Wageningen, the Netherlands VictorM. Toledo Centra de Investigation en Ecosistemas,Nationa l University ofMexico ,Apdo .41- H Sta. Mar/a Guido, Morelia, Michoacan 58090, Mexico Andrew K.Skidmore ITC, PO Box 6, 7500A A Enschede, the Netherlands Jochem Verrelst Wageningen University, Laboratory for Geo-information Science and Remote Sensing, Droevendaalsesteeg 3,Buildin g 100+101, 6708 PB Wageningen, the Netherlands Martijn Slot Resource Ecology Group, Department of Environmental Sciences, Wageningen University, Bornsesteeg 69, 6708 PD Wageningen, the Netherlands Gabriela Schaepman-Strub Wageningen University, Nature Conservation and Plant Ecology Group,Bornsestee g 69, Building 119, 6708 PD Wageningen, the Netherlands 235 Ignas M.A. Heitkonig Resource Ecology Group, Department of Environmental Sciences, Wageningen University, Bornsesteeg 69,670 8 PD Wageningen, the Netherlands GarryBakker Resource Ecology Group, Department of Environmental Sciences, Wageningen University, Bornsesteeg 69,670 8 PD Wageningen, The Netherlands Reinoud Vermoolen Resource Ecology Group, Department of Environmental Sciences, Wageningen University, Bornsesteeg 69,670 8 PD Wageningen, The Netherlands Samuel Lopez de Aquino Museo de Zoologia "Alfonso L. Herrera", Facultad de Ciencias UNAM, Apartado Postal 70-399,Mexic o D. F., Mexico 236 PE&RC PhDEducatio n Statement Form With the educational activities listed below the PhD candidate has complied with the educational requirements setb y the C.T. de Wit Graduate School for Production Ecology and Resource Conservation (PE&RC) which comprises of a minimum total of 22 credits (= 32 ECTS = 22 weeks of activities) Review of Literature (4 credits) - Spatial complexity in agro-ecological systems: bird diversity and coffee plantations. Metacommunity model (2001) Writing of Project Proposal (5 credits) - Birds,traditiona l coffee plantations and spatial complexity: the diversity puzzle (2001) Post-Graduate Courses (11.5 credits) - Computer management of ecological data (2001) - Sub(tropical) land use (2001) - Advanced Statistics (2005) - Multivariate Analysis (2005) - Landscape ecology: its methods and their applications (2005) - Community Ecology (2005) Deficiency, Refresh, Brush-up and General Courses (2.8 credits) - Scientific writing - Techniques for writing and presenting a scientific paper (2002) - Basic statistics (2005) PhD Discussion Groups (3 credits) - Statistics, maths and modelling inproductio n ecology and resource conservation (2002/2005) PE&RC Annual Meetings, Seminars and Introduction Days (0.75 credits) PE&RC annual meeting "Food Insecurity" (2001) PE&RC annual meeting "Biological Disasters" (2004) PE&RC annual meeting "The truth of Science" (2005) International Symposia, Workshops and Conferences (3.4 credits) - IX Simposio, la investigation y el desarrollo tecnologico, Mexico - XX Congreso de Etnobiologia, Univ. Aut. Chapingo 237 - Symposium "Trends in geo-information' (2004) - Master class "state-of-the artremot e sensing in agro-ecosystems (2004) - Landscape ecology in the Mediterranean: Inside and outside approaches IALE (2005) Laboratory Training and Working Visits (2 credits) - GIS and remote sensingprocessing , ITC, The Netherlands 238 •-—•S* S &b R 0 R R ft. •2 .1 s <$ 3 s r-1 r**> '—( £• s 6 0 a<3 So R •a 1 53 I 0 ^1 p .g L *> O ft. £ R 0b •SJ H •J -Si OH CL, OH u .60 0 05 nJ 00 H 00 H 5 ft, -J OH HJ 0, Discussion